This repository has been archived by the owner on Apr 4, 2018. It is now read-only.
aws/gce: Standardise LB health check configs #48
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
For all load balancer health checks on AWS and GCE. Using variables so that they are always the same. Most of these values have been reduced because it was taking a very long time for new instances to come into service. Change the following: - interval to 5s which is the minimum supported by AWS. This has reduced AWS from 30s and increased GCE from 1s. - timeout to 2s which is the minimum supported by AWS. This has reduced AWS from 5s and increased GCE from 1s. - healthy threshold to 2 requests. This has not changed AWS or GCE. - unhealthy threshold to 2 requests. This has changed AWS from 10 and not changed GCE. I'm not 100% confident about the values. They weren't thoroughly tested when we first introduced them for GCE and I suspect we might want to experiment with them in the future, but this is a good start. I've changed the target for API on AWS from the default of `TCP:8080` to `HTTP:8080/info` in order to match GCE and give a more accurate check. Other targets remain as their defaults but we have to pass them because they're mandatory. They don't match GCE because GCE can't do TCP health checks. This will *not* apply cleanly to existing GCE environments due to a bug in Terraform. This should be fixed in the future by hashicorp/terraform#1894. But for the timebeing I think it's important enough that we should delete existing forwarding rules, target pools, and health checks, then let Terraform recreate them with the correct config.
dcarley
force-pushed
the
standardise_health_checks
branch
from
dcb7de6 to
caaec5e
Compare
|
👍 this is great, I like the pulling out of the variables and making it consistent particularly. I agree that it's fine just to start with these values and see how we get on. How do you suggest we merge/roll this out given that it won't apply cleanly? |
I guess merge and let people sort out their own environments if they have one that they can't destroy right now. Contradicts what I said in another PR and I don't think we'd want to do this regularly, but now is a better time than any while we don't have a shared environment. |
|
OK. I'll email to alert team. |
|
Haven't heard this is a problem for anyone, so merging. |
annashipman
added a commit
that referenced
this pull request
May 14, 2015
aws/gce: Standardise LB health check configs
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
For all load balancer health checks on AWS and GCE. Using variables so that
they are always the same. Most of these values have been reduced because it
was taking a very long time for new instances to come into service.
Change the following:
from 30s and increased GCE from 1s.
from 5s and increased GCE from 1s.
changed GCE.
I'm not 100% confident about the values. They weren't thoroughly tested when
we first introduced them for GCE and I suspect we might want to experiment
with them in the future, but this is a good start.
I've changed the target for API on AWS from the default of
TCP:8080toHTTP:8080/infoin order to match GCE and give a more accurate check. Othertargets remain as their defaults but we have to pass them because they're
mandatory. They don't match GCE because GCE can't do TCP health checks.
This will not apply cleanly to existing GCE environments due to a bug in
Terraform. This should be fixed in the future by hashicorp/terraform#1894.
But for the timebeing I think it's important enough that we should delete
existing forwarding rules, target pools, and health checks, then let
Terraform recreate them with the correct config.