We are pleased to announce the release of the Snowplow Collector version 2.7.0.
This new release adds just one new optional feature, which should be considered experimental but potentially very helpful.
Warm-up feature (experimental)
Background: We’ve found that in some production setups the collector can return malformed http responses for a brief period after it first starts up. If the collector is behind a load balancer then these appear to the end user as HTTP 502 errors. This especially happens in Kubernetes deployments, where the collector experiences a very fast ramp up from no traffic to suddenly very high event volumes.
The newly added feature, when enabled, sends some “warm-up” requests to the collector’s own /health endpoint when the collector first starts up. We have found from experiment this can cut down the number of HTTP 502 errors returned from a load balancer in front of the collector in Kubernetes deployments.
This feature is experimental and might be dropped or modified in the future, hence it’s disabled by default. If you are experiencing the issue in question, by all means enable this feature and feel free to share the results on Discourse.
The warm-up feature can be enabled by setting collector.experimental.warmup.enable = true in the config file. See the example config files for a demonstration of where to put this configuration option.
Upgrading to 2.7.0
If you are already running a recent version of the collector, then upgrading to 2.7.0 is as simple as pulling the newer docker image, without making any other configuration change:
docker pull snowplow/scala-stream-collector-kinesis:2.7.0
docker pull snowplow/scala-stream-collector-pubsub:2.7.0
docker pull snowplow/scala-stream-collector-kafka:2.7.0
Check out our documentation pages for a full guide on running and configuring the collector.