-
Notifications
You must be signed in to change notification settings - Fork 267
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
router stopped while loading #2712
Comments
could you explain what you mean by "failed to start" here? Is there a panic? Is it hanging? |
From the log file, it looks like the router is shutting down normally: It's not obvious, however, what is causing the shutdown to occur. I'll make a guess though... I note (from the logs) that the router is shutting down approximately 30 seconds after it starts. This is probably a health or readiness problem in a kubernetes deployment. |
if I understand it correctly the process exit after failing to process a request. you can see this from the attached log file: |
healthcheck issue might seems the issue. "status":404 ->>> it is unclear why this fail with 404 has something changed? |
@talr1 Things have changed, see the CHANGELOG.md for (many) details. The router has a health_check built-in: https://www.apollographql.com/docs/router/configuration/health-checks/ So, that's the recommended way to check that the router is healthy. Our helm chart has Values.yaml that contains configuration for this:
If you add this to your configuration router.yaml, then you should be good. |
Thanks for the prompt help. end of log looks like:
`supergraph: headers: include_subgraph_errors: health_check: and the k8s deployment health to :
logs file: |
It looks like you are using So, just pick which one is your preference. I think |
nice and quick catch. sorry for this typo. |
Ok, so now it seems to work fine, thanks a lot for the help. if I may another question, I've tried adding otel telemetry (was the whole point of upgrading to 1.11.0 due to fixes in otel)
from my testing it seems that the amount of spans are tied to the router log_level (APOLLO_ROUTER_LOG). with level 'info' i can see ~8-9 spans as listed above. with level 'warn' no spans are emitted. I haven't checked other levels. IMO: there should be at least 1 span on each service when i look on a full trace regardless of the service log level. in fact tracing is configured on the router but none is emitted in warn level. Thanks in advance for any help, |
I'm unclear if this was actually an issue, but since the suggestion to use the provided router health_check has resolved the problem I'm happy to close this. @talr1 I'm going to migrate your last section to a discussion question and link it back to this issue. |
Describe the bug
We've had a v1.7.0 router setup working on k8s. when we tried to move using v1.8.0-v.1.11.0, they all failed during router startup. no heavy load issue here. just simply start the router container.
attaching TRACE log
To Reproduce
Steps to reproduce the behavior:
--hot-reload --config router.yaml --supergraph supergraph.graphql
Expected behavior
router should work as in v1.7.0
Output
If applicable, add output to help explain your problem.
router_1_11_0_log.txt
router.yaml:
supergraph: path: /graphql listen: "0.0.0.0:4000" introspection: true headers: all: request: - propagate: matching: ".*" include_subgraph_errors: all: true
The text was updated successfully, but these errors were encountered: