Sevag H

Alright, @connesc - I'm back and ready to work on this again but I'm now realizing I don't like the solution I implemented. I liked pq the way it was...

If I change the code of resample.rs in this repository's examples to resample the two_vowels.wav clip from 44100 -> 44100, I don't have the same problem. I wonder if it's...

Looks like after the `metric[]` slice is constructed and encapsulated in the BulkIndexRequest `sample`, it's never freed/released?

Inuse space good: ``` shanssian:prometheus-es-adapter $ go tool pprof -inuse_space -top http://bleep:9000/debug/pprof/heap | head -n10 Fetching profile over HTTP from http://bleep:9000/debug/pprof/heap Saved profile in /home/shanssian/pprof/pprof.prometheus-es-adapter.alloc_objects.alloc_space.inuse_objects.inuse_space.039.pb.gz File: prometheus-es-adapter Type: inuse_space Time:...

Inuse_space for good measure (since I may be misunderstanding alloc_space):  Also, runtime.MemStats: ``` # runtime.MemStats # Alloc = 163905888 # TotalAlloc = 197824212744 # Sys = 1208749816 # Lookups...

I'm still monitoring the situation but it looks like adding `ES_BATCH_MAX_AGE=30` is preventing this issue. Perhaps it's related to the GC running?

I've been trying to look at the code since my previously deployed version of this adapter (which still works fine). The only thing I can see, since your wrapper around...

I've actually switched to using metricbeat and the Prometheus module. You can use metricbeat to scrape the same targets as Prometheus, rather than go through the `remote_*` adapters.

I was running it with the default settings. Seems like it felt a bottleneck with elasticsearch, and the Prometheus remote storage queue created too many shards/goroutines. Today, I ran it...

Before: ``` ExecStart=/usr/local/bin/prometheus-es-adapter \ --es_url=https://myelastic:9200 \ --es_user=local_prometheus-adapter \ --listen 0.0.0.0:9201 ``` I didn't modify the default values for: ``` ES_WORKERS | 0 | Number of batch workers ES_BATCH_COUNT | 1000...