IBM / Instana Flashcards
(287 cards)
1
Q
How to uninstall and reinstall the agent?
A
- To debug this at your cluster, you have to use a kubectl version which match the cluster kubernetes-api.
-Then you could run the following commands to delete the Agent and install the Agent via your normal approach:
helm uninstall instana-agent -n instana-agent
-kubectl delete ns instana-agent
-kubectl apply -f PATH_TO_YAML
-wait until you see CrashLoopBackOff (I am using watch ‘kubectl get pods -n instana-agent’ to watch the namespace for changes) - Get the events for the pod with status CrashLoopBackOff via kubectl get event -n instana-agent –field-selector involvedObject.name=instana-agent-REPLACE_WITH_CORRECT_IDENTIFIER
-Also get the output from the describe command of this pod kubectl describe pod instana-agent-REPLACE_WITH_CORRECT_IDENTIFIER -n instana-agent
2
Q
Where should you disable python autotracing?
A
In com.Instana.plugin.python section in the agent’s configuration.yaml file, you can disable the python autotrace for that agent. Either use an exclude that matches your application or disable the python autotrace for that agent.
3
Q
Are gunicorn or uvicorn supported by Instana auto tracing?
A
No gunivorn support was removed because recent versions of gunicorn cannot reliably reload middleware once started. This can work once the Instana restores support for gunivorn via the webhook is ready.
And Uvicorn is meant to be used with a manual wrapper.
4
Q
Context Guide
A
It is a GPS that shows you how the infrastructure works together with the architecture and business logic and you will get only the relevant data.
5
Q
What does agent do?
A
Agent automates issue detection, discover of technology, metric collection, and tracing.
The agent is the hub where agents deliver telemetry and trace info to.
Instana agent uses multiple sensors (there is a Java sensor) that initiates a discovery data collection and checks every 20 seconds
6
Q
For node what do you need to install?
A
One line of code
7
Q
What is the fastest cadence to see spikes?
A
1 second timeframe
8
Q
What does the back-end of the agent do?
A
It collects data, does its calculations, and correlate dependencies among applications, services, and end points.
9
Q
How are Instana tenants usually partitioned?
A
Tenants are usually partitioned into three units: production, training, staging - all of which reside in the same tenant.
10
Q
Can all of the tenant units spare the same config?
A
Yes.
11
Q
What settings can each of the tenant units be configured with different settings?
A
Different user permissions, agents, group memberships, and applications.
12
Q
How to deploy agent from Instana Dashboard
A
Click Deploy Agent to walk through a dialogue on deploying an agent based on the environment and install a one-liner.
13
Q
How often does Instana release new features?
A
Every two weeks
14
Q
Where can you find new Instana releases and read about old releases?
A
Http://www.Instana.com/docs/releases
15
Q
How does the dynamic agent check for updates and what is the exception deployed in China?
A
Once we push a dynamic update to a stable branch, the agent will check the repo for new updates by 4:30am if there are new updates. It does this for all units across the tenants.
In China they use a static agent, which comes with all of the info and doesn’t dynamically download from the repo because if they did there would be latency and bandwidth issues from China downloading from repos residing in Europe.
16
Q
Can you favorite each section of the Instana dashboard, can you share it?
A
Yes, you can favorite it and highlight on application perspective and share it
17
Q
What can you see on the Infrastructure page and how can users organize in groups and what is the auto organization setting?
A
On the Infrastructure page, you will see hosts monitored by Instana.
User have put hosts in zones but you can override the zones.
Instana auto groups it by server zone.
18
Q
How can you customize the Infrastructure UI by containers?
A
Lower right-hand corner to change the views. You can zoom in or zoom out.
19
Q
What does red, yellow, or green show on the infrastructure ui?
A
Red, yellow, or green show health of the host/container when Instana backend has over 200 health rules.
20
Q
How do you search for the host/container?
A
You can search for it on the comparison table that by default shows all of your hosts.
21
Q
What is the second column of the comparison table?
A
Click on the second column and then you will be brought to a dashboard for that host and see what agents are collecting and measuring for you.
22
Q
What do agents collect and Instana visualize on dashboard?
A
It will collect and visualize data on context switches, cpu load, etc. and you will see what you are actually running with that host, so called tags, data on network interfaces, Amazon EC2 instances, spring boot application, etc.
23
Q
What does a time slicer do on a graph and what does this help you do?
A
Use the mouse to get all highlights on all of the chart drop down or zoom to timeline. This helps you correlate the same time on all graphs and you can share intervals of a graph with a short form url.
24
Q
What two ways do you find Kubernetes data?
A
To find the Kubernetes namespace page, You can see label, spec, and pods, memory resource, annotations, events, etc.
Second way is click on the Kubernetes icon on the left hand side and you will see all of your Kubernetes clusters. For each Kubernetes dashboard, there is a stack button and upstream/downstream button.
25
What does the stack button do and upstream/downstream button do?
Instana gives you relevant full stack overview for this cluster and you can see all of the application and calls there.
For upstream/downstream, it will show you who is sending calls to this cluster?
26
What do you type in the search bar to find processes for garbage collection?
Type in entity.type.jvm in the search bar and it will find processes for garage collection.
27
Besides search bar, how else do you find the garbage collection section?
If you click on spring boot application, there is a garbage collection section.
28
How does one know what we support?
You you go to the supported technologies page on Instana, it will show you what frameworks, platforms, tracing, sensors, jvms, libraries, which metrics we support.
29
How is an “application” defined?
An application is defined as always dynamic and defined by the user.
Application is the 3 circle icon.
30
What is services on Instana?
Services are what Instana automatically collects and they are based on a library of rules.
*Service names are derived via Instana by the predefined rules
31
What can you do on the predefined rules page?
On the predefined rules page, if it will hit one of these predefined rules, it will stop processing the data and start applying the predefined rule for that service.
Service names are derived via Instana by the predefined rules
32
How to find out how to Install Instana?
Go to the “Install Instana” page on Instana’s documentation.
33
Why self hosted for Apple?
For security and hipa reasons, They want to hoard and lock their data not send their own data. They don’t want our agents to send data outside of their environment.
34
Do we need to set up mounts?
No, customers do and customers need to setup mount points and external volumes:
- Data stores (defaults to /mnt/data
- Cassandra, should not utilize the same volume for production installs (defaults to /mnt/metrics)
-Clickhouse, should not utilize the same volume for production installs (defaults to /mnt/traces)
-Logs (defaults to /var/log/instana). If required you can create a different directory.
35
How is tenant units structured in a url?
{unit}-{tenant}
36
What is the lowest level of endpoints?
Database
37
How to create a license?
Go to main dashboard and go to upper right-hand corner and click on “Create License” and select “NFR-Free License”.
38
What does the Instana 3 Step wizard for application perspectives do?
*Instana has a 3 step wizard for an application perspective. Advanced mode allows you to pick tags, which can be a specific call or params. You can also define level of app like no downstream is jvm level and nothing further, then pick top option. If you want downstream db and messaging service, then pick middle option. For all hubs from there, pick last option.
39
What can you configure application perspective to do?
You can configure application perspective that shows only calls from a source.
And You can change the flow maps on the top so that it does it for In order for Instana to detect downstream services.
40
What are the issues with tracing or tracking Kafka?
*Sometimes you see tech going into Kafka but not going out and this happens when Instana doesn’t monitor this tech service. Or this can also happen when the trace going into Kafka is not correlated with the trace going out. A good question is to ask if Instana can figure out if these calls belongs to this trace.
41
What happens to the Infrastructure ui once you restart the hosts?
The “infrastructure” sub-tab will show all containers start up simultaneously if the host restarts
42
Log messages sub tab
*There are “log messages” sub-tab. You can look up all calls with a particular error message. Same idea with ‘error messages’.
43
Error message sub-tab
In the sub “error messages” tab, if you click on a sub-error, then you will see all the calls causing this error message. It goes to the analytics page. The analytics icon is the magnifying glass icon. Each call is correlate to a user accessing that call. The trace was a post and delete statement. Then you find more logs of the stack trace in the lower right hand corner. We can also see what the profiling is of this call and see cpu, memory, wait time hotspot every few minutes by technology profiling. You can deep dive into metrics by technology service. For Java, it is always on. Try the flame graph
44
Why disable certain default events when Kubernetes is decommissioned?
*If you are running in kubernetes, kubernetes will spin up a new pod. But if not, Instana won’t automatically know it is decommissioned. You can disable certain default events
45
What does system rules->offline and detection do?
System rules->offline and detection will check if entity is still there if host go offline.
You can filter out for technology process criteria under system rules.
46
What do you on the Settings->Alert page?
On Settings->alert page, you can assign events to each alert
47
How many ways can you send an alert?
• *They send web hook, slack, email, etc. an alert can send for multiple events
48
What is a scope for an alert?
scope is how often alerts will trigger
49
What does the agent management page do?
Agent Management page you can view memory and cpu resources of the agent
50
• *Triangle icon
Events
51
•*How do you search calls by latency?
Analytics->Applications-> Calls: you can search calls by latency
52
How is to create alert button
• On a service page, there is a create alert button
53
What is the advanced mode to configure alerts?
You can also set advance mode to configure alert to average performance across some or all calls and services. If your application overnight has less traffic, you have the ability to config alerts less often and reduce the false positives.
54
What are the training resources?
Http://www.instana.com/training-resources
55
To instrument inside the method for Instana
To instrument inside the method, you can either add annotation inside the method and in the code and have Instana agent, like in the yaml file, that defines the method/class name and capture the return value and parameters.
The agent can also capture the stack trace
56
What should you put in a yaml file?
If there is a class within a class, you need to define it in the yaml file.
yaml is a hard deployment where if you change anything you have to restart agent on production.
You can add/edit entry, span, and edit types in yaml.
Apple engineers call this load testing at the method level.
In the future, Ted said instana engineers are trying to build a sensor to auto detect this (what?) so we would have to use a config yaml file
57
Where do you find the analyze->call feature and what does it do?
• On service page, there is Analyze Calls button that brings to Calls page grouped by endpoint. If you filter this page for browser, you can group by end point
58
Is there a doc for configuration for Java trace?
There is an Instana doc for “configuration for Java trace’”.No shotgun or over instrumentation like at Dynatrace
59
What does it mean that Instana can’t do constructors for Java?
• Can’t do for constructors for Java
60
Why does Instana link everything to the infrastructure?
We link everything to the infrastructure that is running the call so makes it easier for us to track the root cause analysis
61
How do you learn more about custom http with Instana?
• *There is a Custom HTTP Headers instana document
62
Does Instana support open jdk?
*We are supporting open jdK.
63
When was Instana founded and why was it founded?
• Instana was started in 2015. Our founder, Mirko, was working with working with our competitors that there were great difficulties getting the current visibility tools to work and the current problems didn’t solve the problem of visibility into dynamic applications
• They noticed that people want to do things quickly. Before, you had to stop multiple times on a Kansan sprint to monitor
64
What does the dynamic graph do?
• Dynamic graph pin underpins Instana. A lot of things can go wrong and we want to get you to the root cause as soon as possible
65
What do sensors do?
• The sensors digest and collect data, then it sees if data meets auto health rules, and you can set your own. Based on info, it decides if we are going to send out events
66
Which 4 type of metrics do we collect and how is Instana different?
We collect these 4 types of metrics: metrics, events, configuration, traces. A lot of other app tools do not collect info on configuration tool
67
How come there isn’t data retention beyond a year?
• *We are discussing internally if we are gonna do a life contract. We notice that people don’t look beyond a year
68
What does infrastructure map use?
• *Our infrastructure map uses open gl
69
What does the time picker do?
• Time picker is sticky. We will have a release tab as well. You have a calendar picker at the bottom
70
Can you test agents locally?
• You can install instana agents manually and test locally
71
What two values does the agent need for the license and where does it go?
Sales key and agent key and it goes into the HCL file
72
What is the difference between an on prem license and a saas license?
When creating a saas license from portal, it goes directly to the dashboard. When creating an on prem license, it gives me a license
73
How do you telnet an Instana on google cloud?
Gc_shh ted_feyler Ted robotshop Telnet the external ip/onpremip 80
If you are running an onprem backend, it is
Telnet localhost 80
Whenever you are in the Linux server of the cloud instance you need to go to /opt/Instana for aws and do a “telnet localhost 80” command or do a “telnet localhost 443”.
74
You can edit /etc/host file for what?
To ping a name rather than an ip address
75
How to install Kubernetes
Different ways like operator, helm chart, yaml, etc.
76
What is helm?
It is a package manager for Kubernetes and half of our customers already use helm and 40% use yaml, operator is the new cool kids
77
Why yaml over helm?
Yaml has everything in one file and helm splits it across different charts
78
On GCP, what do you need to run an agent?
Stand up a virtual machine and a cluster on gcp
79
What to enter on gcp form to create a cluster and how many nodes does it come with and do you pick autopilot or standard mode?
Cluster name, zone
It comes with 3 nodes
Standard mode is good
80
How to get command line script for gcp and do you need to install anything first?
Checkbox the cluster and right-click and hit “Connect”.
Install the gcloud command first
81
What does Kubernetes install on your computer and in which directory?
It installs a ./kube directory in your home computer and inside ./kube directory it installs cache/, config, and http-cache folders and these folders run Kubernetes
82
How to run kubectl
Install kubectl
Cd into .kube
Kubectl get pods
You need to point at right cluster by
“Kubectl get pods -n Instana-agent”
83
What is another name for yaml?
It is daemonset
84
How to apply yaml file?
Kubectl apply -f Instana-agent.yaml
85
Whenever you should edit the yaml file what should you do or else
It will fail?
Kubectl delete -f Instana-agent.yaml
Then
Kubectl apply -f Instana-agent.yaml
86
What is the command to get pods?
kubectl get pods -n instana-agent automatically
87
What command will allow you to list all of the yaml files?
l *.yaml
88
What happens on the dashboard after you delete the yaml?
The cluster name also is deleted
89
How does deploying Instana on Kubernetes make people’s lives easier
One way is you have to run only one command like installing the agent and it does this for 500 nodes
90
What port does robotshop run locally if not kubernetes?
Port 8080
91
What command can you find external ip address that is listening to port 8080 for robot shop and what do you do for the external ip?
Kubectl get svc web -n robot-shop
It is http://externalip.com
92
How do you edit the Kubernetes web definition document to double check what port it is listening on?
Kubectl edit svc web -n robot-shop
93
What is Instana’s elevator pitch?
Instana is AI powered APM solution that automatically monitors containerized micro service applications and solves the problem of visibility into dynamic application
94
What are 6 key concepts of Instana?
-Auto Discovery
-1 second granularity
-100 percent tracing
-Comprehensive dependency graph
-Easy to focus & filter
-Machine learning to automate issue detection
95
How does automated discovery process of all technology works?
You deploy one agent for all languages and it will auto detect technologies in that host or cluster and then reaches out to a remote Instana repo to download and deploy sensors but the sensors have different specialities like the mongodb sensor, nodejs sensor. Every 4am, Instana agent will reach out to the remote repo to see if there are updates to the sensor.
96
How do you find out the sensors each Instana agent has downloaded?
Go to Manage Agents and then click on sensors
97
What type of sensors do each languages has downloaded?
-Automatic metric collection sensor at 1 second granularity
-Automatic tracing
98
Which 4 languages don’t have automatic tracing sensors and what do you do if customers are missing tracing?
Nodejs
Golang
Ruby
Crystal
Point them to documentation
99
Dynamic graph
Instana’s internal model that continuously correlates dependencies between traces in the code and the infrastructure “. Instana’s entire ui is built on the graph.
100
What are the boxes on infrastructure page and what do the height of the boxes represent?
They represent hosts and the height of the box represents horsepower
101
What is the new and better way to navigate through the dynamic graph in the Instana ui?
It is the Context guide or the grey stack and upstream downstream buttons. In drop-down you can drill down through all of the entities and correlated metrics
102
How to use telnet
Telnet {url or localhost} (port number)
103
How does Instana reduce footprint in network transmission?
There are some optimization techniques on metric side. We don’t send the same value in every time. If your host cpu value, we hit 10 percent. If it is still 10 percent we don’t send a new value. If it goes to 11 percent we won’t send in a new value.
104
Can users configure the built in events at Instana and how do you find the built in health rules?
No and there is Instana documentation about the built in health rules
You can workaround it and disable built in events and create a custom event around your unique thresholfd
105
What is a simplified way of capturing metrics vs. traces?
Metrics capture for infrastructure and traces capture for application on top of infrastructure
106
What is 5-prong metric data retention policy of Instana?
1 second granularity is stored for last 24 hours. We have 100 percent of traces in last 24 hours up to last 7 days.
5 seconds for 24 hours
60 seconds for 1 month
5 min for 3 months
1 hour for 13 months
107
What is the traces data retention for Instana?
100 percent full traces for last 7 days
Significant calls and kpis for 1 or 13 months or more.
We keep the significant not the fast traces. All customers should focus on the last 7 days of traces but nothing is guaranteed.
108
How long do you keep dynamic graph configuration and events configuration?
Life of contract
109
What is a call?
A call described an activity within a monitored process, typically a request between two services
110
What is a trace?
A trace is imposed of one or more calls
111
Yes does Instana do distributed traces?
We trace different apps across different hosts and clusters
112
When are traces analyzed?
In real time
113
How does Instana know what the services and endpoints are?
They are automatically discovered
And relationships to other services, endpoints, and infrastructure are automatically correlated and stored in the dynamic graph
114
When are kpis for thoughpit, latency and error rate calculated?
On the fly
115
What do application perspectives do?
They provide filtering and grouping of services
116
What are endpoints?
Endpoints alllow a fine grained view into what operations the services expose.
Every endpoint has a single, automatically discovered type like batch, database, http, messaging, rpc, graphql, etc.
Endpoints can be based on call tags like a combination of {call.http.method} {call.http.path} would be a typical endpoint of a http service
117
What is a service?
A service is a bucket/ collection of endpoints that typically map to infrastructure
118
What is a path template?
/product/{sku}
119
What should be provisioned for robotshop?
2 cpu and 8gb of ram
T2.large
120
What port does robot shop listen to?
8080 under Custom TCP
121
What does Ted recommend to talk to your EC2 instance?
Ssh using .pem file
122
What command to use to find host name of your Amazon box?
Hostname
123
Where does Instana drop the files once you install the Instana agent?
In /opt/instana
124
What is in the /opt/Instana/agent/bin folder ?
The commands
You can list inside the /bin/ directory and run ./status
125
What is in the /opt/Instana/agent/etc folder?
Where config files are stored
126
What is in the /opt/Instana/agent/etc/Instana folder?
It is where you configure the Instana agent.
The most important document is configuration.yaml
127
In the comparison table, how to filter for zone?
Entity.type=“Robot Shop Training”.
128
What command will allow you to look for a keyword like “java” in the process in the local file system?
ps -ef | grep java
129
What does Instana do to traces so that application perspectives work?
It tags each trace
130
What is immediate downstream calls?
They are direct downstream calls on the flow map
131
When creating application perspective what should you pick under “Store Calls of Downstream Services”?
Most of the time pick All Downstream Services
132
Does application perspective get historical data?
No, any change does it for tracking data, moving forward
133
Instana licensing
One agent per 10 containers
134
How can customers make tags for application perspective?
They can push values into call.tags (call level) by instrumenting their code (which means trace) through our Instana sdk
They can tag the agent on the infrastructure page’s entity pop up card and go to “All filters” on application perspective and select agent.tag parameter
They can tag the call.http.header on the infrastructure page’s entity pop up card and go to “All filters” on application perspective and select agent.tag parameter
However, Instana auto tags and customer can add filter
135
Can different services share the same endpoints?
No, services can have different endpoints but endpoints can’t share the same services
136
What does application perspective create?
Generated dashboards
Filtered services
Filtered issued and incidents
Corresponding infrastructure
Drill into analytics
137
What does the upstream/downstream show you?
The call trace
138
What else can you enter in the comparison table search bar?
Entity.type:Java
Entity.jvm.version:8
Entity.host.fdqn
event.severity:warning
139
What does Instana favor between Dapper and Open Tracing?
OpenTracing
140
What is Instana most compatible with?With tracing frameworks?
Otel, zipkin, Yayga.
We offer support for opentelemetry metrics
141
How do you start using opentracing?
You can send opentracing spans to Instana.
In Instana config file you open the endpoint to open telemetry
142
What is preserved by the headers
The context of Instana traces are preserved by the headers.
143
What is W3c trace standards?
W3c standard is enabling operableness between different systems with different header ids.
144
Why do you add a trace header context with open tracing, and can you have two different trace libraries from two different applications going into one unified trace.
Add a trace header to manipulate the header as well.
145
What is network time?
Network time is the time between services
146
What happens if you try to setup something unique with endpoints?
Instana places a limit on this and tried to prevent endpoint explosion
147
How can you rename endpoints?
Go to Configure Endpoints to rename endpoints
148
What can you do to endpoints if you have a bunch of unique ids?
Group them
149
How to Rename service names?
You can use Instana documentation on predefined rules and you can change the service rules here in the training environment at services/configure/new and it tells the Instana pipeline to define a predefined rule to change the service name.
150
How does Andy define events?
A chance/change? is what agent discovers every 20 seconds
Issues are id there is a negative health change
Incidents are triggered by edge services
There are list of online and offline events
151
How do you see what causes the issue, event, incident?
Click on View Built-In Event to see what caused it.
When you click on the issue, it will show you the triggering event.
152
What is a maintenance window?
It allows you to avoid alerts during particular times
153
What happens when you switch off to “incident”?
You can specify grace period
154
What happens if you set offline event detection to yourself?
You can set an alert to yourself,
155
Custom metrics?
Streaming micro meters through Prometheus remote write
156
How can you push and pull events programmatically?
Using Instana’s agent rest api
157
What does error and warning logs try to do?
It looks at thresholds of historical logs and recommends thresholds for you
158
What do smart alerts try to auto do?
It looks at historical data and recommends a static or historical baseline for you.
There is also a number of blueprints like these subtabs on left hand corner.
You can adjust the sensitivity.
It versions the alerts.
159
Custom dashboard and how to tap into Grafana plugin.
Create your own custom dashboard with widgets.
There is a Grafana plugin for Instana and you need to create api token in Instana.
In upper left hand drop down go to settings->api tokens
160
What does exclamation point and fed or yellow triangles on Timeline box indicate?
Indicates errors
161
For Instana, what does butler api write to and what is butler api for?
Butler writes to cockroach db
162
What does the Instana backend filler do and how does the infrastructure mapping break due to filler and how to fix it?
We have a component called a filler which takes data and puts it in data repositories. When we hit the filler limit, the service mapping issue was breaking. We needed the service to map to infrastructure. Where the infrastructure mapping breaks is wheee the service mapping breaks. We know we can remains the filler limit and make adjustments on the backend.
163
What did we have for elastic search for apple mobile events and what happens when you adjust the snapshot cache in filler?
Whenever we adjust the snapshot cache in filler, that affects elastic search which is running a little bit because is only a 4 node cluster. JP added one node for the elastic search cluster to get bare level nodes but we didn’t need to do this for apple events.
We have 3 nodes for elastic search that are 264 gig ram and 32 cpu hosts.
164
What do we typically need to do to adjust/increase filler?
Increase heap for filler change.
165
Does Instana handle nested spans?
Yes we use trace id to stitch all spans together.
166
Which box can you install Robotshop?
When you install your Instana backend, you don’t want to install the Instana backend and instal robot shop on the same box. The backend to emulate customers’ installing an agent on their host should be a space different from the space for robot shop.
167
If you did “docker compose down”
it will turn down the containers.
168
When telnet command said “telnet unable to connect to remote host:Connection refused”
- do a “Audi Docker ps” command to see the processes and container names that are running one
169
What do you need to do before getting the install command line going?
You need to do a “Instana unit” again before you do “Instana —help”, and you need to do “Instana start” then “Instana unit”.
170
What is the first thing we do when the metrics or tracing comes in?
First thing we do is check it against the health rules. Let’s say the cpu comes in at 99 percent instead of 10 percent then it is unhealthy.
171
How were standard size in profiles defined in for Apple?
They were defined in yaml for apple.
JP uses a different heap size for each profile and JP can provide these profiles, which cpu and ram used by dropwizard apps, which we use to process the data.
172
How do you pick sizing of profiles?
JP days it makes it easier if the person needs to provide a baseline. For example if you want to do metrics that means you don’t have a lot of application data. If you know the expected thoroughputs and entities that you are monitoring they will help us determine sizing
173
For apple, what happens if we get sizing wrong at the beginning? Can we size it up or down?
JO is moving profiles from yaml config to single settings hcl. Let us know when we need to change the size and our sre team will change it for apple
174
On which port are our load balances open to the world on?
Port 443
175
How does Instana monitor msk, rds, and amazon keyspaces?
Ted said you need to deploy an agent so you are gonna have agents running on application obviously and collecting metrics and traces. For Amazon services like msk and rds, we deploy another agent on EC2 instance that is configured as an aws agent and collects info through cloudwatch. Ted says amazon supports keyspaces but we don’t currently support it.
176
Apple asked for next practice on tagging
If you are leveraging amazon ec2 tag and have that amazon agent running, then we collect or correlate all of the tags you already created. Ted says we don’t have any best practices except for a key value tag pair.
177
Apple asked how does Instana manage permissions?
Ted said within Instana you have a roles based access control system tbst is pretty granular in that it gives you the ability to have users do a lot of things except for infra. Ted days our goal is to use saml for apple and we do the saml as we do the backends. But we are dependent on apple to do tbat and that causes more areas of slow down. Ted says we don’t know if we have ldap groups but we are not sure if it is in there yet. Ted days we are doing saml groups but saml groups probably not supported yet but you can set up permissions at the group level.
178
Where is spark on?
Spark is on emr and emr is the equivalent of Hadoop cluster. They use emr on ec2
179
How many agents per node?
One agent per node. Each agent has its own cpu consumption.
180
Does an agent have its own cpu consumption or does it share cpu consumption with other tenant units?
Yes each agent has its own cpu consumption.
181
Why do we keep a local scope for sensors?
The sensors know nothing about connections to the agent and the agent has the burden of collecting the data, processing it, and kicking it to the back end for further processing and this is how the automatic discovery process works.
182
How fast does Instana telemetry stream?
Telemetry streams in continuously at 1 second resolution to monitor full stack of a service which is very fine grained telemetry to monitor a stack.
183
How to see messages published to the queue or processed to the queue?
You can click on Analyze->Calls on one of these services and the in the message queue transport and you can see the messages published to the queue or processed to the queue and if we supported Kafka it will show the same way.
184
How to use Instana to track where teams make different network calls to where?
Sandro’s hypothesis is to see all of the calls being made by the cluster and look for deviations and isolate the slower calls and then drill into the infrastructure of where the call is slow when it executed and see where from the host down to infrastructure it got slower. So if you find out it is running on a server we shouldn’t be calling that is my red herring.
185
Do we support Amazon MSK or Amazon EMR and how to implement msk with Instana?
For MSK we have some support for emr in that we have a sensor that collects telemetry for emr and this feature will show up as an infrastructure component and is becoming part of the infrastructure. Emr shows up as a service in our infrastructure map and cloudwatch needs to be configured and the agent needs to be installed. Not every component of infrastructure is hardware and it can be a service on the cloud.
The sensor here is a proxy to cloudwatch to collect metrics and then we start correlating traces. In this case we track only infrastructure metrics and we need to be able to instrument the run time.
Sandro recommended implementing the Java trace sdk.
186
If you have 4 different aws accounts, how do you pull cloudwatch data from these many accounts?
You use emr
187
How to see from application level to database level?
Select db row on analyze calls and you will see the sql query on the right
188
What is Palumi?
Palumbo is a provisioning for infrastructure framework
189
How many entities and host does an Instana backend handle?
Each backend handles a hundred thousand entities and 20,000 hosts.
Most of the time with apple we had one business unit per backend.
190
A trace is defined as
A series of calls and traces comprise of spans at the lowest levels
191
What is synthetic monitoring Instana?
It projects what real users are doing. The problem with real users is that they aren’t using the app and they simulate by having synthetic or fake users hitting the app at certain intervals or constantly. They will run a script from the local computer and they can run it from machines in different areas. It is not real user monitoring like google analytics
-it tracks availability of an endpoint
-provides repeatable performance monitoring in tracking api results
-validates complex api workflows and data that is being displayed
192
Ela?
Enterprise license agreement
193
What takes the most most cpu: remote write and prometheus or statsd and prometheus?
Supposedly remote write
194
What is the max number of spans we have done for apple events?
80-90k spans which is 20 hosts in apm mode. Nate fisher has load tested it to have 300-850k spans.
195
Apple’s kube is
MicroVM based
196
How does JP optimize?
Increase snapshot limit to 50k
Heap size increased by 2GB
JP sets the HEAP_OPTS value to 22GB and two minimum memory values to 30GB or 32GB. Will James said 14 ad a value for CPU is good and HEAP_OPTS are parameters that come ijnKafkan& Cassandra libraries. We Instana don’t play around the HEAP OPTs too much.
JP made sure all of data stores are running in same data center as these components. And when he runs elastic search, it is a multi node and distributed. Everything that requires a quorum has at least 3 nodes. JP will look into whether or not xxx/large is the largest profile for this version and that’s why they have overwrites. He change UI backend to 1GB.
Will James says that UI backend profile has xxl-profile has 17 and 22 for memory. Maybe we just do 12 for memory. We can’t scale ui backend horizontally, just vertically and it depends on the size of state is storing and because it is doing a lot of garbage collection.
There was a proxy for the network that apple that might be contributing.
JP will create a script that will auto create the new load balancer and it will the glcb (global load balancer) and point that at the new load balancer in front of the acceptor. The new UI will be faster because there are two load balancer, one for the acceptor and one for ingress and this should speed it up because will be one less hop because current hop is slow. .
197
Can you scale good components of hardware vertically or horizontallly?
Both if you have good hardware.
There are limits depending on load
198
Is it harder to scale of components on SaaS or on-prem with Instana?
You can scale filler out on on-prem but 10x scalability project should resolve it. Different code makes SAaS scale filler out better than the on-prem code.
Fillers do run in a scalable config and some of the SaaS units run with 150k entities with a single filler instance and can scale out no more than 200k entities. When apple wanted 35k entities but on-prem needs to specify some parameters in code and this is going to require code changes to get this configured Corte fly and our team is working on it.
In SaaS none of the TUs is no bigger than 200k with a singer filler.
With a single JVM, 35k is no problem but we can scale it up to 100k for on prem, hopefully not it depends on the type of workload depending on the metrics going into the Instana system. But if your infra is dynamic then scaling couldn’t help here. Apple wants 100k over 35k.
199
What is filler?
Filler is a processing component for pumping data in the db
200
How did we propose to fix. The time infomap ui for apple?
The time the ui takes to load doesn’t have anything to do with filler. If we scale up the ui backend vertically, it can improve the load time of the informal ui, giving it more heap and cpu, and things like that.
Also fixing the hanging web sockets that talk to the ui backends. There is a tcp transmission error on ui backend and tcp packet keeps getting dropped
201
How did apple go into the Kubernetes node?
Vi ~/users/thaidi/.bash_profile
Cat ~/.bash_profile
eks —region us-west-2 update-kubeconfig —name pos-dev-us-west-2
kubectl get namespaces
kubectl get pods -n Instana-agent
(Look at image pullback)
kubectl describe pod Instana-agent-string -n instana-agent
Kubectl apply -f instana_new.yaml -n
Kubectl get pods -n instana-agent
Kubectl describe pod Instana-agent-string -n instana-agent history
Kubectl logs -f instana-agent-string -c instana-agent -n instana-agent
202
What does the error “livened probe failed”?
Perhaps this is hosted in the apple data center where there is no internet there but this means they need to host the static image somewhere.
203
What are the Docker commands to pull the latest Docker image that apple used when their data center had no internet?
Docker pull Instana/agent:latest
Docker images | grep Instana
Docker tag Instana/agent {url of docker}
Docker push {url of docker}
204
How to pull the latest image?
Go to the “Packages (DEB, RPM)” drop-down value. A url shows up and grab that URL.
It goes to a portal url at packages.instana.io:443
Username is _
Password is the agent key
Grab that agent key in between the “:” and the //
#download the static image, powered by the Azul JDK command
Docker pull containers.instana.io/instana/release/agent/static:latest
Docker login containers.instana.ui -u _ -p
Docker tag containers.instana.io/instana/release/agent/static:latest
Docker push {apple docker url}
205
What happens when you pick the left handed sidebar for Amazon dc2 and choose “Elastic Computing (EC2) - Linux”?
Vi instana.sh
Sh instana.sh
Then the command is asking for the password
Cat Instana.sh
Chmod 755 instana.sh
206
Where did we put the logs for apple’s ec2?
/opt/instana/agent/data/log
207
For apple, how many traces were we capped at for application perspectives?
We weren’t capped at 500,000-800,000 traces that we scaled up to. 16,000 nodes ideal. There are a lot of constraints for how application perspective scales.
1m entities gets to 10,000 nodes.
We are not going to get to a point where we get 1 backend and there is 1 million hosts in it so we are going to have to split the host across multiple backends. We don’t have a built in way in the product to federate the backends. Potentially you can have a trace split across them and it will be the same trace id
208
Why does and why doesn’t Nate fisher thing the trace graph is a huge issue?
NF doesn’t believe the trace graph is a huge issue but he thinks the problem is the traces don’t get recorded unless we receive the root spans. So we need to make sure we record trace’s regardless of whether or not we receive the root spans.
209
Why is there a Bunch of work required to make tea dd across Tenant units?
Nate fisher conjectured if we captured traces as they came and we did a scatter gather query across TUs or sub section of TUs, then maybe we can surface that trace info at a high level. Doing the slicing and dicing across Tenant units is problematic. To find a specific trace, probably won’t be an issue.
210
What has NF noted about an issue in communicating across multiple platforms pertaining to Kubernetes?
The issue is as soon as you start shipping data between the jvm sensors and the agents, it creates all sorts of issues because there’s no way to apply a standard policy because it requires dynamic port allocation and bidirectional communication.
For things like service metrics, we went around this by booking around ip tables creating bypasses.
But until we actually rework that communication there is no way we can use proper policy definitions and that’s shF apple is going to require in their custom cmi plugin for Kubernetes.
But apple has their own custom network plugin. If it is what NF thinks it is, NF didn’t think we have a way to work around it because if it is a programmable network then the only way we can work is having it as cloud native, implying there is a fixed number of ports that we connected g on and the sensors are connecting down to the agent, or we are connecting over a service port for the given service and we are dialing into their exposed service over a service port.
211
Is custom metrics inexpensive?
Custom metrics is interest because from an operational standpoint it is more of additional components and that increases costs and custom metrics are 2-3 times more expensive to operate. It will cost them millions per month for custom metrics.
212
How to filter for endpoints?
Go to Analytics > Application/Calls page and then filter by Call Name = POST /odata/v9/getfailedlogonattemptforuser group by endpoint name.
And you will see two endpoints of post/odata and post/odata/9 and then you can look up the last two hours or last 24 hours.
213
Should you uncheck the two endpoints such as post/odata and post/odata/9 - in the “Edit Global Smart Alert” box?
Because if you uncheck these two boxes in the Edit Smart Alert box, it will work but it will be too broad and it will remove all of the calls of this endpoints and overfilter (quirk of Instana).
Alternatively, leave all of the boxes “Edit Smart Alert” box and add a filter towards the bottom of the “Edit Global Smart Alert” box that filters for Call Name != POST /odata/v9/getfailedlogonattemptforuser
There is the other option which is to create a specific endpoint for that call and that call and that checkbox for that specific endpoint will appear in the tree selection and the result will be the same.
214
How does Instana automatically discover endpoints?
Each service comprises of endpoints which are automatically discovered by Instana using endpoint rules and are labeled by Instana to have either a path template and if there is no path template, it takes the first part of the url so Calls are longer.
215
How to Configure Endpoints for /odata/v9/getfailedlogonattemptforuser
On the service page, click on “Configure Endpoints”. Add a custom http file and paste odata/v9/getfailedlogonattemptforuser in and then test it. If you put the literal string, it will be broken down by its endpoints using instana rules.
Alternatively you can use variables like in /odata/v9/(action) and deliberately controlled that all of those calls will be broken out in their own individual endpoints.
You can also configure endpoints using the Instana rest api
216
Why is creating a custom endpoint better than adding a filter even though both ways would produce the same results?
Creating a custom endpoint allows you to see more visibility and creates more value on your dashboards and saves you time from drilling further. Another reason is that you can put variables in parameters like /odata/(version)/getfailedlogonattemptforuser It will change only data moving forward, not backwards. Takes longer than a half hour for data to percolate moving forward.
217
How does Yahoo Japan do custom metrics and can custom metrics be built in metrics?
There are 2 different frameworks like Micrometer and Prometheus
And they capture micrometer metrics at the JVM level:
Entity.JVM.micrometer.gauge:
So at https://training.Instana.instana.io there is micrometer custom metrics at the JVM
Custom metrics aren’t built in metrics
And custom metrics will appear on the host dashboard
218
How does it work for custom infrastructure metrics?
You can create a custom widget for custom infrastructure metrics and this exists for the Instana SaaS product.
Not side of this exists for the on-prem pricing but if it doesn’t exist, you have to configure it with grafana
219
How does one query for arlo’s Prometheus metrics?
entity.jvm.micrometer.gauge: > 0 && entity.jvm.app.name:”Instana Demo - Discount Application 1.0”
Entity.Prometheus.gauge: > 0
Entity.Prometheus.counter: > 0
If we look at Arlo prod - https://prod.arlo.Instana.io they have Prometheus metics in the custom dashboard so type “Entity.Prometheus.counter: > 0”
Pick “Infrastructure & Platforms” for the “Data Source:” drop down
220
How to find agent logs programmatically?
cd /opt/Instana/agent/
cd bin/
/.status
cd data/
ls
cd log/
tail -f agent.log
(If issues and make sure you are in the bin directory)
Sudo ./stop
Sudo rm ../data/log/agent.log
Sudo ./start
Sudo ./status
Ls
Sudo rm ../agent.pid
Sudo rm ../lock
Sudo ./start
Tail -f ../data/log/agent.log
Sudo ./status
221
What are webhooks?
Webhooks are user-defined http callbacks. For example, a service developer can provide an http/(s) url to third-party service and ask them to notify on certain events.
Third-party service calls the provided url with a certain payload to convey information about the events.
222
With Apple, what happens if you get the error message, “Connectivity with one or more of the backend”?
Cd etc/
Cd Instana/
Sudo cat com.Instana.agent.main.sender.backend.cfg
*ec2 cannot connect to port 443 because maybe due to firewall and if that is the case, ping won’t help
223
How to download all of the sensors for a host in the Instana dashboard?
If you click on the hostname link in Instana, please scroll all the way down and click “Open Agent Management” button and then click on “Sensor Info” or “Download logs” button and you’ll get all of the sensors.
224
How to change the entity type to Kafka and entity type to msk?
Cd etc/instana
Vi configuration.yaml
Sudo vi configuration.yaml
In the yaml file, adjust to:
entity.type: kafka
entity.type: msk
*we don’t know if msk was added before 195 or after 195
225
Can you have more than one tenant per backend and how is data storage configured for the backend?
Data storage is defined by and per backend.
1 backend can have multiple tenants.
In apple case, backend storage wasn’t configurable by apple.
226
What is data retention policy for 1 second data?
Considered short-term data. 1 second data is kept for 7 days.
227
What is the data retention for an event and alert?
It is considered long term data and it goes up to 1 year
228
Where are short term data and long term data stored on Instana’s side?
They are in different storage centers so short term data location gets shipped to the long term location and they talk to each other.
229
For Instana and Apple, csm backend A see backend B?
If you are in backend A, you can’t see the data from backend B. If you are in backend b, you can’t see the data from backend A.
230
Can you define more than one agent to send data to a backend?
Yes you can define more than one agent to send data to one backend
231
Why do we care if the agent can ping 443 endpoint?
As long as the agent pings 443 endpoint, Instana is able to ingest the data and we don’t care where the data is coming from
232
What works for 99% of people installing Instana agent on Kubernetes?
JP said Daemon set YAML is 99% of people that would install the agent on the eks cluster on the eks cluster on the installing Instana-agents page. JP uses helm to install it.
233
If you are having issues getting python sensor to work what do you do?
You need to “Configure Network Access for Monitored Application”. There is a documentation.
If you add the python variable to the container, it should start monitoring the python application. You don’t have to add any particular sdk.
234
Where to manage logs in the Instana ui and what does JP use?
Go to the gear sign in the UI.
JP personally uses the LogDNA logs -‘s pulls logs from AWS.
235
How do you monitor Amazon Cloudwatch?
Monitoring Amazon Cloudwatch requires manual configuration and it isn’t an automated process.
236
What does eks stand for?
Elastic Kubernetes Services
237
What are cross ganaway’s 3 synthetic test types that will come first before browser testing.
We plan to support https endpoint monitoring by accepting the token, http monitoring, simple ping. And these three would come first before rolling out browser testing.
You can expect Instana to have the ability to store credentials like passwords, global variables, and api tokens in a vault that is accessible to script programming.
238
What does cross ganaway expect for the browser testing feature when Instana rolls it out?
For browsee testing expect customers will be able to record for browser testing wherever the tests are located.
-You’ll be able to store system info as name/value pairs and user defined information like stage environment like if this is a dev, staging, production environment.
-you can jump start by recording browser reactions by starting with a test script.
-cross wants to combine synthetic monitoring with unbounded analytics for basic api and scripted api tests which provided end to end visibility with the infrastructure visibility
239
What is cross ganaway’s intended power of end user monitoring for scripted browser testing?
Distributed traces show the end to end flow and end to end website visibility
240
How is Cross Ganaway going to display synthetic monitoring results?
In custom dashboards and all integrated with Instana alerting:
-system defined incidents
-user defined incidents
-sli/slo
Wants to do this with public facing apis and websites and results are sent to the Instana SaaS through public apis while you have Instana agent installled locally
241
Regarding synthetic monitoring, can customers run the agent on a different cloud other the customer’s application is on?
Results are being sent to Instana SaaS from the other cloud using public apis. We haven’t determined the best locations to deploy.
242
What did Costco express that they want?
Costco definitely wants to https pinging and eventually would like to monitor someone being logged in and make sure all of the screens show. They are now doing only browser testing
243
What to do if one of the pods is giving ErrImagePull under the “Status” column?
First check to see if the developer is able to access any repo outside of their security network.
Second, check if you hit the Docker limit. Docker has a registered limit on how many images you can pull every few days. There might be an administrator at the customer’s company who pushes images to the Docker url.
Third, JP recommended to switch to the Instana repo- container.instana.ip - so that way you run the Daemon set you’ll always get the latest image and you don’t want to bother anyone. You will need to login to containers.instana.up and I need to pull and edit the yaml into daemon set and you need the use the Instana agent key.
244
How to subscribe to Instana status?
You can subscribe to status.instana.io
245
What kind of events do we have for when Java goes offline? And what are the system rules for Java? Is Java supported by Instana?
We have offline events or types of events or custom events whenever the Java agent goes offline.
There are two system rules, one for host and another for beyond the host like the JVM.
Instana auto-captures Java application and log information.
246
How to configure a custom event for springboot app?
1) go to settings->events->new events
1. Set “Source:” to “System Rules”.
2. Set “System Rule” to “Hosts do not have matching entities running on them”.
3. Set “Entity type” to “Spring Boot App”.
4. Set “Entity Label Operator” to “contains”.
5. Set “Entity Label” to “Ends 1.0” entity. Springboot is an entity.
6. Set “Offline for” to “5 min”.
7. Set “3.Scope” to have one of those three things.
A. All available entities, an option you don’t want to pick because you’ll get overalerted.
B. Application perspective, which means you pick one of these Applications under the Applications page.
C. Dynamic Focus Query language, which you use in the comparison table “entity.hostname.{domain url}.com” OR “entity.hostname.{domain url}2.com”
247
How do you configure custom event for file system usage over 90 percent for a specific host name?
You can configure a custom Event at Settings->Events->New Events for File Systems > 90%:
Set “Source:” to “Built-in Metrics”.
Set “Entity Type” to “Host”.
Set “Metric” to “Filesystem: Used”.
Set “Matching Operator” to “is” if you want to type in the absolute path or Set “Matching Operator” to “contains” if you want to contain a specific word.
Right now, you have 5 hosts on Instana’s comparison table with entity.host.name=*mqmft*, so therefore your hosts with entity.host.name=*psapp* hasn’t been scoped with a corresponding custom event alert. I believe you have 6 psapp hosts.
You already have a “CP4I_Host_FSUsage_W” custom event that alerts when “any” filesystem which is scoped to a particular host of entity.host.name=*mqmft* and that is at or above 90% usage for a time window of 10 minutes. You can modify this existing custom event alert with Dynamic Focus Query under “Scope:” to have entity.host.name=*mqmft* OR entity.host.name=*psapp*, or you can create a new custom event alert from scratch.
It’s recommended that you use the section under “Scope:” to scope the custom event alert to a particular host or particular drive/relative url, because you don’t want to create a custom event alert for every drive and host.
Please always test your AND OR operator arguments in the “Infrastructure Map” feature as shown in the attached video
248
How do you create a garbage collection alert?
You can create a garbage collection alert by:
Go to “Alert Channels”->”New Alert Channels” to Create an alert channel for each or both of two email addresses.
Then go to “Alerts”->”New Alert” to create an alert and select “Alert on Events’ option in the dropdown.
Under the “Events:” section, please click on the “Add Events” button and a popup box will appear. In the filter input box, search for and click on the “Garbage collector activity high” event.
Then select under the “Alert Channels:” section, select the Alert Channels by clicking on “Add Alert Channels”.
It’s recommended that you use the section under “Scope:” to scope the custom event alert to a particular host or particular drive/relative url, because you don’t want to create a custom event alert for every drive and host.
Click on the “Add Events” button at the bottom pane to submit your new event alert
249
What is first way to create a Java offline alert?
There are 3 ways to create an Java offline alert by:
Option 1:
Go to “Alert Channels”->”New Alert Channels” to Create an alert channel for each or both of two email addresses.
Then go to “Alerts”->”New Alert” to create an alert and select “Alert on Event Type”’ option in the dropdown. Then activate the “Change” option to green.
Then select under the “Alert Channels:” section, select the Alert Channels by clicking on “Add Alert Channels”.
Under “Scope:” , you will scope the alert to a particular host or particular drive/relative url. For example, you can scope it to entity.host.name=*mqmft* OR entity.host.name=*psapp*.
Please always test your AND OR operator arguments in the “Infrastructure Map” feature as shown in the attached video.
Click on the “Create” button at the bottom pane to submit your new alert
250
What is second way to create Java offline alert?
Option 2: However this option doesn’t allow you to trigger an alert after the event has been offline for a certain time. The default might be for 30 seconds.
You can configure a custom Event at Settings->Events->New Events for the 6 JVMs:
Set “Source:” to “System Rules”.
Set “System Rules” to “Offline event detection”.
Under “Scope:” , you will scope the alert to a particular host or particular drive/relative url. For example, you can scope it to “entity.host.name=*prextptlbord* AND entity.selfType:java”so you can scope the 6 JVMS that you want to scope.
Please always test your AND OR operator arguments in the “Infrastructure Map” feature as shown in the attached video
251
What is a third way to create a Java offline alert by?
Option 3:
You can configure a custom Event at Settings->Events->New Events for the 6 JVMs:
Set “Source:” to “System Rules”.
Set “System Rules” to “Hosts that do not have matching entities running on them”.
Set “entity” to “JVM”.
Set “Entity Label Operator” to “contains”.
Set “Entity Label” to “Java”.
Select the desired time value for “Offline for”.
Under “Scope:” , you will scope the alert to a particular host or particular drive/relative url. For example, you can scope it to “entity.host.name=*prextptlbord* AND entity.selfType:java”so you can scope the 6 JVMS that you want to scope.
Please always test your AND OR operator arguments in the “Infrastructure Map” feature as shown in the video.
252
What is the correct protocol used for Instana to send spans/traces from agent to Instana?
It is http/2
253
Static alerts and static threshold
Fixed or timed based alerts
254
Adaptive threshold and adaptive alerts
Alerts based on What is current value and what is expected
Adaptive means learned over time and shows how much you can expect after learned over time
255
Global smart alerts
Allow you to create alerts across multiple applications
256
What difference in levels for smart levels? 
Warning level vs. critical level
257
What trigger points do you have for smart alert?
Slow calls, erroneous calls, logs, http status codes, throughput?
258
Why pick inbound calls for smart alerts?
In case you are receiving calls from a specific customer
259
Why would you use filter for smart alerts?
Remove or add services, endpoints, or tags
260
What is unique when you select adaptive threshold?
There is a sensitivity dial. If you select low, then it means a lot of issues happen and an alert will happen within 16 standard deviations. With high sensitivity, you have to get an alert immediately.
261
Persistence over time feature with smart alerts?
If “evaluation granularity” is 20 min, and “number of consecutive metric violations” is 1 and “number of metric violations” is 1, you will get it in 20 minutes. But if you choose 3, then it is 3 x 20 min x 1, you should expect it in 60 min
262
What is trace impact with smart alerts?
If you select 20 for “# of impacted traces” then this means you will get an alert if 20 of your metrics are impacted
263
What does “insert placeholder” do in Alert Channels?l
You can insert dynamic information in the alert message for that alert channel
264
What can you adjust when creating a smart alert in an alert channel?
1) insert placeholder in title or description
2) select warning or critical
3) turn on triggers incident on or off
4) write description
265
Custom payloads
They are custom key value pair info that you can pass along with your alerts.
266
What are 3 different ways to building our custom dashboards?
-Build native dashboard within Instana.
-Leverage our Grafana plugin to combine Instana data with other data sources?
-Query our Web test API and build dashboards in the tool of choice
267
What data sources can you build your custom dashboards off of?
Applications (Traces and Calls)
Events
Infrastructures & Platforms
Mobile Apps (beacon)
Websites (Beacon)
268
How do you create a new dashboard?
The “Instana” logo in upper left hand side has a drop down and click on Add Widget.
269
Shortlink lasts as long as?
You can share this visualization and if you share it through short link it becomes a custom event that is stored as long as The duration of the event in the data retention period
Share button allows you to change the custom dashboard to be private or public
270
What are two types of smart alerts?
Application perspective smart alerts
Website & monitoring smart alerts
271
What are Instana key performance indicators and how do you find smart alerts from the application tab?
Go to Applications->Smart Alerts
(1) latency
(2) number of calls
(3) errors
272
Include Internal calls in relation to a smart alert?
Internal calls are work done inside a service
273
Include synthetic calls in smart alerts?
Health checks are handled by synthetic endpoints in micro services like Kubernetes
274
How many entities and applications perspectives can you apply to a global smart alert?
Many
275
Evaluation granularity for smart alerts means with number of violations?
Means it will evaluate after your chosen threshold of every over 5, 10 min only if 1 or 3 chosen violations occur, we will alert
2nd radio dial of consecutive values: Means it will evaluate after your chosen threshold of every over 5, 10 min only if 1 or 3 chosen consecutive violations occur AND your chosen how many times the consecutive actions, we will alert
276
What is difference between regular event and incidents?
Regular events is when things get unhealthy, and an event is an entity on its own
Incident means access by your users being impacted
277
What is the difference if I create a smart alert from application perspective page?
There is a wizard and it will pick up the chosen application perspective and you can add a dynamic query filter based on that chosen application perspective
When you switch to Advance mode it is back to the normal events->smart alerts page
Different thresholds will appear based on the performance for that application perspective based on the 7 days vs 24 hour timeframe
278
When would you use static type of drop-down value of weekly seasonality for smart alerts?
Say you are a Netflix app and your thresholds will need to be different because user behavior is different between mon vs. fri
279
For smart alerts, when will daily seasonality start kicking in?
When it has learned from data for at least 5 days but it is better if it is 7
280
What is cool about what shows up on application perspective page at the bottom of each box?
They have red icons and these Red icons that tell you when a smart alert is triggered
281
Say you have a smart alert alert already configured and then you add a new application perspective, will the smart alert auto apply to the new application perspective?
No, it doesn’t you will have to manually configure that this new application perspective applies to this new smart alert. But if you add new services and endpoints to an already existing application perspective that is already applied to this smart alert, then yes those endpoints and services will auto be applied to these smart alerts
282
Why are smart alerts better than standard alerts?
Mainly we provide intelligent default recommendations and we are allowing all types of users to create smart alerts
283
Old style vs. new style deploying Instana of k8 and why move to new style?
Old style deploys everything in a daemon set, new style gives option to deploy the k8 sensor as a deployment which makes the rest of the agent (daemonset) slimmer.
When you have issues with k8 discover and more than 10 nodes, we strongly suggest to go to the new deployment if you haven’t.
284
Why it is recommended to have an internal cache for dynamic agent?
Dynamic agents upgrade every night and we recommend to have an internal sensor cache so not each agent goes out to the internet but only to the internal sensor cache (nexus)which hardens the nightly updates.
285
How to search /opt/Instana/agent/bin/
Do ./status
Ps -ef | grep instana
Ps -ef | grep java
286
Built-in event for Instana agent self-monitoring
Go to Events left-hand sidebar, then pick “Built-in Metric” under Source.
“Entity Type” is “Instana Agent” and “Time Window” is 30 seconds
“Metric” is “Network > Sent” with “”avg” is “Aggregation” and “s” is under “Operator” and 1 is under “Bytes”
287
How to do self-monitoring on the agent as a custom event?
Go to Events left-hand sidebar, then pick “Built-in Metric” under Source.
“Entity Type” is “Instana Agent” and “Time Window” is 30 seconds
“Metric” is “Network > Sent” with “”avg” is “Aggregation” and “s” is under “Operator” and 1 is under “Bytes”
“Scope” is “Apply on (required)” and “Selected Entities Only (Dynamic Focus Query)
Under “Dynamic Focus Query:” put “entity.agent.mode:APM OR entity.agent.mode:Infrastructure”
