Report generation fails for big projects

[nehashri]

Report generation fails if there is a huge project due to plugin connection timeout. The current default value for gauge to send a kill message to plugin is 4 secs. For a huge project this timeout is not sufficient to generate reports. Few ways to address this:

• Stream results to html report (instead of passing the entire execution result)
• html-report sends back a "keep-alive" ping, asking gauge to extend the timeout.
• ask user to use a "saner" timeout value per project.

(decreasing order of complexity, also decreasing order of cleanliness (IMO))

[v-mwalk]

nehashri - what was the resolution with this? My project adds a significant quantity of logging to the results and hence the 4000mS timeout is hit. Reckon your first thought (Streaming) would be the nicest way - although, as you point out, the most complex.

[sriv]

@v-mwalk - some relevant discussions on this issue

We will implement the second approach (sending ping requests) and thus letting the plugin live for longer as opposed to enforcing a hard timeout. Does that sound ok?

[gaugebot[bot]]

The fix should be available in nightly >= 2-4-2019

[gaugebot[bot]]

The fix should be available in nightly >= 10-4-2019