I'm trying to aggregate values from multiple metrics into one (obviously) but that's not happening and I'm desperately trying to figure out why. There's not much to go on though and I'm hoping get some more information on how to debug this.
Let's start with the environment:
Debian 9.11
graphite-carbon 0.9.15-1
The contents of my aggregation-rules.conf are:
app.all.values.<metric> (1800) = sum app.*.values.<metric>
My storage-schemas.conf has a section for these metrics:
[app]
pattern = ^app\.
retentions = 15m:15d,30m:30d,1h:180d,4h:2y,8h:4y,1d:8y
I do not have a section for app inside storage-aggregation.conf. It is my understanding that this isn't necessary?
The data is sent to graphite every 900 seconds (15 min) and I can see and query all the individual series just fine directly on graphite or from a grafana instance.
When I start a graphite-aggregator instance, I get these lines in the aggregator.log file:
07/10/2019 17:25:43 :: reading new aggregation rules from /etc/carbon/aggregation-rules.conf
07/10/2019 17:25:43 :: clearing aggregation buffers
But no aggregated metrics are ever created. I did have syntax errors in the beginning, but fixed them. Since then the aggregator process starts and runs successfully, it just never aggregates any metrics.
So how would I go about debugging this?
Do your metric names have only 4 nodes, or is supposed to match across multiple nodes? You may need to use <> to match across more than one node/across dots.
For example, if you have a metric called "app.myapp.values.requests" it would match your existing rule, but "app.myapp.values.requests.count" wouldn't. However if you changed your rule to
app.*.values.<<metrics>>it would match both examples.Does that help?