Deployment & Provisioning Flashcards
InstanceLimitExceeded error
You have reached the limit on the number of instances you can launch in a region.
- 20 by default
InsufficientInstanceCapacity error
AWS does not currently have enough available On-Demand capacity to service your request.
- Wait a few minutes and try again - Request fewer instances - Select a different instance type - Purchase Reserved Instances instead - Submit a new request without specifying the AZ
What happens when you hit the IOPS limit of your volume? How to fix it?
App becomes to slow down and I/O request starts queuing
- gp2: increase the size of your volume (only if volume is < 5.2TB)
- if you have already reached 16,000 IOPS, chage storage class to Provisioned IOPS
Bastion Host
It’s a host located in a Public subnet
- Allows you to connect to your EC2 instances using SSH or RDP.
- Used to securely connect to instances in a Private Subnet, without exposing them to the internet.
- Does not enable outgoing requests, internet access for your instances in the private subnet.
Application Load Balancers
Layer 7
- Inspect packets
- Make routing or routing decisions based on the contents of the network packet.
- acloudguru/sales → goes to one server
- acloudguru/marketing → goes to another server.
- It can read HTTP, HTTPS headers.
Network Load Balancers
- Layer 4
- Load balancing TCP traffic
- Millions of requests per second, while maintaining ultra-low latencies.
Classic Load Balancers
Layer 4 and 7
- X-Forwarded-For, which simply allows you to determine the originating IP address for the device which sent the request.
- Allows Sticky sessions: simply binds a user’s session to a specific instance.
When should you Pre-Warm your Load Balancers?
Needed when there will be a crazy increase in your traffic
- Contact AWS and request them Pre-Warm ELB.
- It will configure the ELB to the appropriate level of capacity based on the traffic you expect.
What does AWS need to know in order to Pre-warm your load balancers? (3 things)
Start and end dates
Expected request rate per second
Total size of a typical request
Which ELB type should you use if there is a need for static IP addresses?
NLB
- It attaches 1 static ip address per subnet
400 ELB error
Bad or malformed request
- e.g. Request header malformed. - doesn't conform with HTTP or HTTPS standard.
401 ELB error
Unauthorized - user access denied.
- The user trying to request the web page, they don’t have access to the web page or the website that they’re trying to access.
403 ELB error
Forbidden
- Request is blocked by WAF access control list
460 ELB error
Client closed connection before the load balancer could respond.
- client timeout period may be too short. - it doesn't give the LB sufficient time to actually respond to your request
463 ELB error
LB has received an X-Forwarded-For request header, with more than 30 IP addresses.
- X-Forwarded-For request header allows you to identify the IP address of the originating requester. - Limit is 30 IPs per request - More than 30 you get a 463 error message.
500 ELB error
Internal server error
- e.g. Error with the LB
- Wrong configuration of the ELB
502 ELB Error
- bad gateway
- e.g. Your application server, it closed the connection or it sent back a malformed request to the load balancer.
503 ELB Error
Service unavailable
- No registered targets - it simply refers to the web servers themselves that are registered with the load balancer
504 ELB Error
Gateway timeout
- e.g. if the application is not responding.
- Maybe you've got a problem with your web server, your application server or database
561 Error Code
Unauthorized
- if your load balancer is using an identity provider to authorize or authenticate users. - if you don't receive a response from the identity provider - something's gone wrong with the identity provider - you can't reach it for whatever reason
What does BackendConnectionErrors shows?
Number of unsuccessful connections to backend instances.
HealthyHostCount
Number of healthy instances registered
UnHealthyHostCount
Number of unhealthy instances
Latency
number of seconds taken for registered instance to respond/connect.
