Merge branch 'main' into update-actions-versions

doc/performance-investigations/0024-enable-minified-html/outputs/baseline/2024-06-20T15:05:18/runner-benchmark-g4u-g55zd_output_failures.csv

@@ -0,0 +1 @@
+Method,Name,Error,Occurrences

doc/performance-investigations/0024-enable-minified-html/outputs/baseline/2024-06-20T15:05:18/summary.yaml

@@ -0,0 +1,14 @@
+---
+Percentile Averages:
+50th: 67ms
+90th: 108ms
+95th: 146ms
+99th: 253ms
+99.9th: 400ms
+---
+GETs (99th): 267ms
+POSTs (99th): 238ms
+---
+Total Requests: 86,409
+Total Failures: 0
+Error Percentage: 0.0%

doc/performance-investigations/0024-enable-minified-html/outputs/investigation/2024-06-20T10:31:13/runner-benchmark-x1f-gj8rr_output_failures.csv

@@ -0,0 +1 @@
+Method,Name,Error,Occurrences

doc/performance-investigations/0024-enable-minified-html/outputs/investigation/2024-06-20T10:31:13/summary.yaml

@@ -0,0 +1,14 @@
+---
+Percentile Averages:
+50th: 66ms
+90th: 107ms
+95th: 145ms
+99th: 262ms
+99.9th: 458ms
+---
+GETs (99th): 274ms
+POSTs (99th): 248ms
+---
+Total Requests: 86,023
+Total Failures: 0
+Error Percentage: 0.0%

doc/performance-investigations/0024-enable-minified-html/outputs/investigation/2024-06-20T10:57:05/runner-benchmark-73g-n8825_output_failures.csv

@@ -0,0 +1,3 @@
+Method,Name,Error,Occurrences
+GET,/questionnaire/sections/calculated-summary-section/{id}/,CatchResponseError('Expected a (200) but got a (0) back when getting page: /questionnaire/sections/calculated-summary-section/cERcgg/'),1
+GET,/questionnaire/sections/section-companies/,CatchResponseError('Expected a (200) but got a (0) back when getting page: /questionnaire/sections/section-companies/'),1

doc/performance-investigations/0024-enable-minified-html/outputs/investigation/2024-06-20T10:57:05/summary.yaml

@@ -0,0 +1,14 @@
+---
+Percentile Averages:
+50th: 68ms
+90th: 108ms
+95th: 144ms
+99th: 258ms
+99.9th: 512ms
+---
+GETs (99th): 265ms
+POSTs (99th): 251ms
+---
+Total Requests: 86,370
+Total Failures: 2
+Error Percentage: 0.0%

doc/performance-investigations/0024-enable-minified-html/outputs/investigation/2024-06-20T11:40:18/runner-benchmark-z5o-s4znr_output_failures.csv

@@ -0,0 +1 @@
+Method,Name,Error,Occurrences

doc/performance-investigations/0024-enable-minified-html/outputs/investigation/2024-06-20T11:40:18/summary.yaml

@@ -0,0 +1,14 @@
+---
+Percentile Averages:
+50th: 65ms
+90th: 106ms
+95th: 145ms
+99th: 257ms
+99.9th: 432ms
+---
+GETs (99th): 261ms
+POSTs (99th): 252ms
+---
+Total Requests: 86,308
+Total Failures: 0
+Error Percentage: 0.0%

doc/performance-investigations/0024-enable-minified-html/summary.md

@@ -0,0 +1,67 @@
+# Enable HTML minification investigation
+
+Minification of HTML was previously turned off due to performance issues before the Census. We need to investigate the
+impact of re-enabling it in Runner.
+
+Following on from a recent investigation of enabling HTML minification in Runner, we concluded that re-enabling
+minification did not have any adverse affects on our
+UI.
+
+However, during the spike, we ran a basic performance benchmark to check if there were any performance impacts. We
+now need to run a comprehensive performance investigation, to check the impact on our response rates and other key
+indicators, such as our CPU usage, when minification is enabled. Enabling minification may have an adverse impact at a
+larger scale.
+
+## Runner settings
+
+| Setting       | Value   |
+|---------------|---------|
+| Concurrency   | 115     |
+| Max instances | 3       |
+| Min instances | 3       |
+| CPU           | 4       |
+| Memory        | 4G      |
+| Version       | v14.3.2 |
+
+## Benchmark profile
+
+Attempted to mimic traffic as closely as possible during a busy period for Runner. To avoid hitting Runner with a lot of
+traffic on a 'cold start' against idle instances, the hatch rate was set to 5.
+
+| Option                 | Value                                 |
+|------------------------|---------------------------------------|
+| Requests file          | requests/test_benchmark_business.json |
+| Run time               | 20m                                   |
+| User wait time minimum | 1s                                    |
+| User wait time maximum | 3s                                    |
+| Clients                | 80                                    |
+| Hatch rate             | 5                                     |
+
+## Results
+
+No errors were present in any of the test runs, so have been excluded from the results
+
+| Metric                                | Baseline | Investigation 1 | Investigation 2 | Investigation 3 |
+|---------------------------------------|----------|-----------------|-----------------|-----------------|
+| Questionnaire GETs (99th percentile)  | 267ms    | 274ms           | 265ms           | 261ms           |
+| Questionnaire POSTs (99th percentile) | 238ms    | 248ms           | 251ms           | 252ms           |
+| Total Requests                        | 86,409   | 86,023          | 86,370          | 86,308          |
+| CPU Usage                             | 16.93%   | 18.61%          | 19.59%          | 19.79%          |
+| Memory Usage                          | 31.98%   | 31.51%          | 31.80%          | 32.12%          |
+
+Performance graphs:
+
+* [baseline](outputs/baseline/performance_graph.png)
+* [investigation](outputs/investigation/performance_graph.png)
+
+## Observations
+
+* Most noteworthy observation here is the CPU usage. Although it's only ~2% higher, enabling HTML minify seems to have a
+  consistently higher CPU rate.
+
+* The memory usage was consistent across all baseline and investigation runs, at ~30%
+
+## Decision
+
+Although negligible, enabling minify HTML seems to have a negative impact on CPU. It also doesn't appear to offer
+any notable performance improvement