Skip to content

LogFlow-AI/metrics-streaming-service

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

16 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Metrics Streaming Service

This is a metrics streaming service that allows us to stream and aggregate metrics effectively!

Developer's Guide

  1. On a terminal window, install and spin up kafka + zookeeper:
# 1. Install Kafka and Zookeeper
brew install kafka
brew install zookeeper

# 2. Start services
brew services start zookeeper
brew services start kafka

# 3. Verify services are running
brew services list | grep -E 'kafka|zookeeper'
  1. Install relevant libraries (Activate custom-environment if needed)
# 1. 
pip install -r requirements.txt
  1. Initialize kafka topic and run the application.
# 1. Delete stale kafka topics
kafka-topics --delete --bootstrap-server localhost:9092 --topic latency-metrics

# 2. Restart Kafka and Zookeeper if needed
brew services restart zookeeper
brew services restart kafka

# 3. Create topic with replication-factor 1 for local testing
kafka-topics --create --bootstrap-server localhost:9092 --replication-factor 1 --partitions 1 --topic latency-metrics

# 4. Verify topic creation
kafka-topics --list --bootstrap-server localhost:9092

# 5. Run application
python ./src/main.py --csv ./data/HDFS_v1/preprocessed/Event_traces.csv
  1. Monitor & Observe
# Terminal 2: Monitor consumer group
watch -n 1 kafka-consumer-groups --bootstrap-server localhost:9092 \
  --describe --group latency-monitor-group

# Terminal 3: Monitor topic metrics (iffy)
watch -n 1 kafka-run-class kafka.tools.GetOffsetShell \
  --broker-list localhost:9092 --topic latency-metrics

Below is an example of the output from the latency monitor dashboard. This was taken from a run of the application on the HDFS_v1 dataset with parallelization enabled (num_workers = 4, partitions = 4).

================================================================================
Metrics Streamer Latency Monitor Dashboard - 2025-05-06 13:31:02
================================================================================

Block ID: blk_-9128742458709757181
Current Latency:   131.00 ms

Performance Metrics:
Messages Processed:   575061
Data Points in Sketch:   575061
Runtime:   462.02 sec
Throughput:  1244.65 msg/sec
Refresh Rate:     0.00 sec

Latency Statistics:
  Mean:     0.00 ms

QuantileFlow DDSketch Percentiles:
  P50:   7260.81 ms
  P95:  50529.42 ms
  P99:  53654.09 ms
================================================================================

About

This is a metrics streaming service that allows us to stream and aggregate metrics effectively!

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages