Delay Tests

The last version (1.3.3) was experiencing major delays when handling a large number of metrics to monitor with a meager step for collecting values. That is the primary improvement yield on the new version of this component (1.4.1) and as thou can see below, the results were substantial, decreasing the delays.

Environment machine specifications: 4 CPU, 8GB RAM, 60GB DISK. Quick note: it is important to mention that depending on the machine that is running MDA, the performance might, or might not, decrease/increase specially when dealing with delays. Also, with, simultaneously, a large number of monitoring specs being posted, there might be some specs failing. But this last use case occurs with exhaustion tests.

---

Figure 1: Delay (min) for scenario without aggregations and 100 metrics (Version 1.3.3)

Figure 2: Delay (min) for scenario without aggregations and 500 metrics (Version 1.3.3)

---

Figure 3: Delay (min) for scenario without aggregations and 100 metrics (Version 1.4.2)

Figure 4: Delay (min) for scenario without aggregations and 500 metrics (Version 1.4.2)

Figure 5: Delay (min) for scenario without aggregations and 1000 metrics (Version 1.4.2)

Figure 6: Delay (min) for scenario without aggregations and 2000 metrics (Version 1.4.2)

Figure 7: Delay (min) for scenario with aggregations and 1000 metrics (Version 1.4.1)

Figure 8: Delay (min) for scenario without aggregations, 10 monitoring specs with 100 metrics each (Version 1.4.1)