No, I'm afraid we don't have polished config testing tools like that yet, alas. We have some ways to approximate what you're asking for, but you have to do it yourself.
The way I'd handle the case you're describing is that I'd set up a Heka instance with the configuration I want to test, except instead of using an ElasticSearchOutput I'd use a FileOutput. When you spin it up, since the FileOutput is using the same encoder that the ES output will be using, the generated file will contain the exact JSON that will be sent to ElasticSearch. We also ship with a simple `heka-inject` (https://hekad.readthedocs.org/en/latest/developing/testing.html#heka-inject) utility that lets you easily generate test messages to make sure that Heka handles them as you expect. Using variations on these themes I've been able to test all of the configs that I've had to develop. Hopefully down the road we'll have tools that will simplify this a bit. -r On 03/18/2015 07:09 AM, Brian Lalor wrote:
I’m considering a switch from Logstash to Heka for all of my log ingestion needs. I like what I’m seeing, but I’m concerned about validating and maintaining my heka configs. Logstash has a really nice test runner that allows me to write very focused tests against pieces of my config. I don’t see such a facility in Heka and it’s making me reluctant to switch. For example, with Logstash, I’m able to write a test that ensures that an event with a particular message and type is properly fed through the appropriate filters and rewritten or modified into a form suitable for indexing in Elasticsearch. That may be grokking something from syslog or adding geoip metadata to an http access log entry. Have folks tackled this yet? -- Brian Lalor [email protected] _______________________________________________ Heka mailing list [email protected] https://mail.mozilla.org/listinfo/heka
_______________________________________________ Heka mailing list [email protected] https://mail.mozilla.org/listinfo/heka
