Chris Martin
Chris Martin
- Migration job for eventdb - Cleanup job for eventdb - Removed caching (queue, jobset) -> ids - Added retry logic for db operations
* Default `eventRetention` in EventIngester to 336 hours (matches the current default in the server) * EventIngester no longer has a hard coded list of retryable redis errors. Instead by...
Add some Prometheus metrics (suitable for alerting) for the lookout ingester. Examples include: * Failed to deserialize messages * Lost connection to pulsar * Error processing messages * Failure to...
Currently the E2E tests don't test any of lookout. Testing the UI might be hard, but we should at least be able to test the state of the database fairly...
Lookout uses the `job.job` database column to display the job yaml. This is a binary json column, which is unfortunately slow to ingest. To solve this, lookout ingester has started...
Add some Prometheus metrics (suitable for alerting) for the event ingester. Examples include: - Failed to deserialize messages - Lost connection to pulsar - Error processing messages - Failure to...
Currently we use Alpine as our base images. Unfortunately, this means that CGO isn't supported which is undesirable. We should research and (if appropriate) recommend what alternatives are out there,...
Traditionally all Armada development has been done by running the entire armada system. As Armada becomes larger and more complex, this becomes increasingly difficult to do. instead, as each subcomponent...
Right now, `ListBuilder, MapBuilder and SortedSetBuilder) are intialised with empty data structures. This is great if you want to create an object from scratch, but doesn't cover the use case...
**Describe the bug** Zookeeper doesn't handle SSL certificate rolling gracefully. Specifically, if a certificate is rolled Zookeeper will continue to use the old, expired cert until it is restarted, which...