Collecting Metrics from Apache Server

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view

Collecting Metrics from Apache Server

Supun Abeysinghe
Hi all,

I'm investigating ways of collecting metrics from the Apache server. Specifically, I'm trying to figure out a way to collect throughput (req/sec) and the average latency. I have tried using mod_status, but it only gives throughput. I'm thinking of using the following approach.

Current Approach:
1. Create a new Log Format including the request processing time.
2. Use Logstatsh to read Apache access logs.
3. Output the timestamp and request processing times to Elasticsearch
4. Query Elasticsearch using a python client and get calculate the performance numbers (average throughput and average latency for the required time windows.)

Is this a good approach to take performance numbers? I'm having doubts about this because, at every iteration of performance calculations, I need to query and get a large number of items (corresponding to requests) from the Elasticsrearch server which I suppose ads a huge overhead.

Any alternative methods or any pointers is highly appreciated.

Best regards,

Supun Abeysinghe
Undergrad, Department of Computer Science and Engineering,
University of Moratuwa, Faculty of Engineering.