Uh oh!
There was an error while loading. Please reload this page.
SAN-5817 - Store container logs in s3 - #1903
Conversation
Updated binding Updated binding Added task Added tail limit Updated to use aws.s3 s -> S Updated to use .then instead of .tap Create instance of S3 Updated to only tail if a tail limit is set Re-working container log storing Updated logging Updated to clean stream Pass end event through Updated trace Updated data log Updated data length Re-ordered code Updated order WIP Being silly Being silly Being silly Meh Push raw data through Go back in time
fix all the stream stuff so we can start saving logs to s3
# Conflicts: # lib/models/rabbitmq/index.js
Myztiq
commented
Mar 1, 2017
| return Promise.try(this._clearBuildResources.bind(this)) | ||
| .bind(this) | ||
| .then(this._updateModelsAndGetUpdatedContextVersions) | ||
| .tap(rabbitMQ.publishContainerLogsStore.bind(rabbitMQ, {containerId: this.id})) |
There was a problem hiding this comment.
() => rabbitMQ.publishContainerLogsStore({containerId: this.id})
There was a problem hiding this comment.
I disagree. Not going to change it unless we've a style guide thing that prevents it.
| .promise() | ||
| .catch((err) => { | ||
| log.error({err}, 'Error uploading logs to s3') | ||
| throw new WorkerStopError('Error uploading logs to s3', {originalError: err}) |
There was a problem hiding this comment.
This is a bit aggressive, might just be network errors. Just add max retry count unless you have specific errors to workerStop
| throw new WorkerStopError('Error uploading logs to s3', {originalError: err}) | ||
| }) | ||
| }) | ||
| .then((uploadData) => { |
There was a problem hiding this comment.
nit: ponos should print out end of worker.
also what is uploadData ? is that the entire log string?!!?
There was a problem hiding this comment.
uploadData is the promise containing the upload results.
There was a problem hiding this comment.
just making sure, is upload results the entire log in memory?
There was a problem hiding this comment.
No, it's the response that we get from s3
Deployed |
Myztiq
commented
Mar 8, 2017
tests passed when I run them in small chunks of under 100 locally on bdd. |
We need to store logs of containers when they die so we can show them for testing environments. We're going to need to add more logic and store container histories so we can fetch old versions etc but this is the first step.
Store them in s3 by container ID, then eventually we'll be able to stream directly to the client from s3 bypassing API sockets.