Benchmarks
2 minute read
Overview
Edge Delta uses the Vector Test Harness to obtain benchmark results.
This tool was developed by Timberio. This tool is an open source tool, located in the https://github.com/timberio/vector-test-harness public code repository.
Test Cases
TCP to Blackhole
In this test case, the flow starts with messages from producers and then terminates with the agent when the messages are ingested to itself.
TCP to TCP
In this test case, the agent ingests messages from TCP producers and then pushes the processed messages to a TCP consumer.
TCP to HTTP
In this test case, the agent ingests messages from TCP producers and then pushes the processed messages to an HTTP consumer.
File to TCP
In this test case, the agent ingests messages from files and then pushes the processed messages to a TCP consumer.
Regex Parsing
In this test case, the agent ingests messages from a TCP, then applies an Apache common regex processor, and then pushes the processed messages to a TCP consumer.
Results
Edge Delta retains the results obtained by Vector and adds results for the Edge Delta agent:
Test | Edge Delta | Vector | Filebeat | FluentBit | FluentD | Logstash | SplunkUF | SplunkHF |
---|---|---|---|---|---|---|---|---|
TCP to Blackhole | 490.3 mib/s | 86mib/s | Not applicable | 64.4mib/s | 27.7mib/s | 40.6mib/s | Not applicable | Not applicable |
File to TCP | 110.2 mib/s | 76.7mib/s | 7.8mib/s | 35mib/s | 26.1mib/s | 3.1mib/s | 40.1mib/s | 39mib/s |
Regex Parsing | 17.1 mib/s | 13.2mib/s | Not applicable | 20.5mib/s | 2.6mib/s | 4.6mib/s | Not applicable | 7.8mib/s |
TCP to HTTP | 211.2 mib/s | 26.7mib/s | Not applicable | 19.6mib/s | <1mib/s | 2.7mib/s | Not applicable | Not applicable |
TCP to TCP | 212.7 mib/s | 69.9mib/s | 5mib/s | 67.1mib/s | 3.9mib/s | 10mib/s | 70.4mib/s | 7.6mib/s |
Vector has already performed benchmark test and published the results for different APM tools, including:
- Vector
- Filebeat
- FluentBit
- FluentD
- Logstash
- Splunk UF
- Splunk HF
For more information, including listed results, see https://github.com/timberio/vector#performance.