Test Properties
| Role |
Protocol |
Test Duration |
Maximum Latency |
API Name |
API Version |
Connections |
Message Size |
Variable Size |
Rate |
Limit Destinations |
Broker URL |
Distribution of the Latency Percentiles
Service Time Statistics
| Metric |
Value |
| Total Count |
N/A |
| Max Latency |
N/A μs |
| 99.99ᵗʰ Percentile |
N/A μs |
| 99.9ᵗʰ Percentile |
N/A μs |
| 99ᵗʰ Percentile |
N/A μs |
| 95ᵗʰ Percentile |
N/A μs |
| 90ᵗʰ Percentile |
N/A μs |
| 50ᵗʰ Percentile |
N/A μs |
Response Time Statistics
| Metric |
Value |
| Total Count |
N/A |
| Max Latency |
N/A μs |
| 99.99ᵗʰ Percentile |
N/A μs |
| 99.9ᵗʰ Percentile |
N/A μs |
| 99ᵗʰ Percentile |
N/A μs |
| 95ᵗʰ Percentile |
N/A μs |
| 90ᵗʰ Percentile |
N/A μs |
| 50ᵗʰ Percentile |
N/A μs |
The latencies displayed in the table above are show in microseconds whereas the
chart displays them in milliseconds.
Tests without a fixed rate will not display response time metrics.
That is expected when running all-out tests because the backend is unable to estimate the response
time based on the rate. Ideally, Maestro performance tests should use fixed rates
and evaluate the pass/fail condition based on the (service time) latency at the 90th percentile.
Rate Distribution Over Time
The rate geometric mean represents the actual average rate obtained during the test. If zero, it means that
at some point during the test, the rate was zero.
Rate Statistics
| Metric |
Value |
| Maximum Rate |
N/A msg/sec |
| Minimum Rate |
N/A msg/sec |
| Mean |
N/A msg/sec |
| Geometric Mean |
N/A msg/sec |
| Standard Deviation |
N/A |
| Rate Samples |
N/A |
| Rate Error Count |
N/A |
| Skipped Rate Records |
N/A |
Test Node Information
| Name |
OS Name |
OS Arch |
OS Version |
OS Other |
HW Name |
HW Model |
HW CPU |
HW CPU Count |
HW RAM |
HW Disk Type |
HW Other |