Starting around the time of the 24Q4 data release, I noticed the public portal slowed down dramatically. Today it’s timing out (see below). The issue doesn’t seem to be localized to my computer or local network / IP address, since I’ve witnessed this with my phone as well. Can you please take a look?
Thank you very much,
Brendan
$date; ping dmc.depmap.org
Mon Jan 27 13:38:15 EST 2025
PING dmc.depmap.org (34.31.22.241): 56 data bytes
64 bytes from 34.31.22.241: icmp_seq=0 ttl=57 time=51.565 ms
64 bytes from 34.31.22.241: icmp_seq=1 ttl=57 time=54.500 ms
64 bytes from 34.31.22.241: icmp_seq=2 ttl=57 time=53.572 ms
64 bytes from 34.31.22.241: icmp_seq=3 ttl=57 time=54.474 ms
64 bytes from 34.31.22.241: icmp_seq=4 ttl=57 time=54.433 ms
^C
— dmc.depmap.org ping statistics —
5 packets transmitted, 5 packets received, 0.0% packet loss
round-trip min/avg/max/stddev = 51.565/53.709/54.500/1.127 ms
$date; ping www.depmap.org
Mon Jan 27 13:38:25 EST 2025
PING depmap.org (23.236.54.134): 56 data bytes
Request timeout for icmp_seq 0
Request timeout for icmp_seq 1
Request timeout for icmp_seq 2
Request timeout for icmp_seq 3
Request timeout for icmp_seq 4
Request timeout for icmp_seq 5
Request timeout for icmp_seq 6
Request timeout for icmp_seq 7
Request timeout for icmp_seq 8
Request timeout for icmp_seq 9
Request timeout for icmp_seq 10
Request timeout for icmp_seq 11
Request timeout for icmp_seq 12
Request timeout for icmp_seq 13
Request timeout for icmp_seq 14
Request timeout for icmp_seq 15
Request timeout for icmp_seq 16
Request timeout for icmp_seq 17
Request timeout for icmp_seq 18
Request timeout for icmp_seq 19
Request timeout for icmp_seq 20
^C
— depmap.org ping statistics —
22 packets transmitted, 0 packets received, 100.0% packet loss
Yes, I had noticed it was slower and had put it into the queue to investigate. I had hoped it wasn’t so slow that it was disruptive, but seems like it is causing disruptions so I’ll prioritize this.
I did some investigation and found two clear performance regressions that were introduced in the 24Q4 update. I’ve deployed a fix for those two issues.
In general, the portal still feels fairly sluggish so I’ll continue looking for any low hanging improvements that we can make. While I don’t feel like the performance issues are fully solved, I believe the changes I made will address the issue that was causing the site to become overwhelmed and non-responsive.
Thank you very much Phil! If it’s helpful, one specific slow behavior I (previously) noticed was the response time from text entry boxes (the gap between typing in a gene or cell line, and the field responding with a dropdown list). It’s usually unnoticeable but can be many seconds now, especially during midday.
When you say, “response time from text entry boxes”, is this a statement about the text input boxes across the portal as a whole, or just data explorer?
(ie: do you feel the top search bar at the top of the pages or other text boxes are also much slower then before, or is it that you’re finding data explorer specifically slower?)
Hey Phil - The portal has been much faster this week, so whatever you’re doing is working - thank you!
The (formerly) slow text boxes were both at the top of the page and in the data explorer. To elaborate - when I typed a search term into text box at the top of DepMap: The Cancer Dependency Map Project at Broad Institute, the portal seemingly runs a database query to return a dropdown of controlled vocabulary terms (so like, I type H1092, then portal helps me find “NCIH1092” as the correct term). The delay was between the user typing in the search term, and the portal returning the dropdown.
In the data explorer, the delay was between selecting a data type (“CRISPR”) and feature type (“Gene”, though I think that gets auto-selected since it’s the only option), and the portal responding with the (“Choose gene…”) dropdown. It seemed like the slow step was the database response, though I can’t really speculate too specifically.
As I wrote though, since your last message the portal has been much much better and I can’t replicate this behavior any more - thank you again - Brendan
edit - I also recall that plots within gene pages were slow to load; you’d get the layout/skeleton of the page but the panels would be empty and take many seconds to populate. Again, this went away a week ago.
