When we tested the portal last week, it was hard to interpret problems without knowing if the underlying services are responsive. Real-time results are needed.
Implement a web dashboard that shows the status of each phylotastic service.
The simplest version of "status" is just "was it running the last time we tested", but this does not mean very much if the last test was a month ago.
Ideally, we would have the response times for small medium and large tasks, for all phylotastic services, relative to the distribution of historical response times. For lists and trees, 20, 200 and 2000 items would be small, medium and large tasks. For image retrieval from EOL, we only want to test small and medium numbers. For PDFs, 1 Mb, 5 Mb, 25 Mb represent small, medium and large.
When we tested the portal last week, it was hard to interpret problems without knowing if the underlying services are responsive. Real-time results are needed.
Implement a web dashboard that shows the status of each phylotastic service.
The simplest version of "status" is just "was it running the last time we tested", but this does not mean very much if the last test was a month ago.
Ideally, we would have the response times for small medium and large tasks, for all phylotastic services, relative to the distribution of historical response times. For lists and trees, 20, 200 and 2000 items would be small, medium and large tasks. For image retrieval from EOL, we only want to test small and medium numbers. For PDFs, 1 Mb, 5 Mb, 25 Mb represent small, medium and large.