Very high memory usage on Windows
I had v1.73.0 installed on a Windows 10 PC. It was running well. I just upgraded to version 1.74.0. The program is running but "hogging" memory on the Windows PC.
The total memory usage is about 83 to 96 percent of the availble memory. The PC has 16 Gb installed. Vmmem is using about 5,500MB, Docker desktop 79 Mb and Docker backend 31MB. The machine is slow to respond to commands
When I rebooted the Windows PC without Docker (and Immich) the PC worked normally. After staring Docker and Immich the memory use went up again.
Is there anything I have done to cause this problem?
How do I cut down the memory useage?
4 Replies
Can you run
docker stats --no-stream? This will show the memory usage for each container.here is the results of docker stats --no-stream
C:\ProgramData\immich-app>docker stats --no-stream
CONTAINER ID NAME CPU % MEM USAGE / LIMIT MEM % NET I/O BLOCK I/O PIDS
08b8388b142f immich_proxy 0.00% 11.71MiB / 12.4GiB 0.09% 198MB / 187MB 0B / 0B 9
d1a619a9defc immich_server 0.02% 183MiB / 12.4GiB 1.44% 171MB / 321MB 0B / 0B 12
21ff9aff529a immich_microservices 0.06% 534.5MiB / 12.4GiB 4.21% 1.53GB / 2.9GB 0B / 0B 12
7a71648739d2 immich_web 0.00% 36.44MiB / 12.4GiB 0.29% 14.1MB / 19.7MB 0B / 0B 12
5eba4aef6723 immich_machine_learning 0.10% 1.986GiB / 12.4GiB 16.02% 14.4kB / 2.43kB 0B / 0B 29
3d3ccf1e18b4 immich_redis 0.27% 19.93MiB / 12.4GiB 0.16% 105MB / 27.3MB 0B / 0B 5
f274926b272d immich_postgres 0.00% 313.4MiB / 12.4GiB 2.47% 148MB / 6.66GB 0B / 0B 7
132fa20be63a immich_typesense 0.12% 860.4MiB / 12.4GiB 6.78% 5.15GB / 22.1MB 0B / 0B 299
What do I have to do to reduce memory usage; current;y the PC is running at 77%
I noticed high memory usage during uploads due to thumbnail generation and image tagging. I had to add about 5GB of memory to my VM to avoid swapping to disk. I’m guessing this is expected behavior though. I did lower the number of concurrent tasks in the settings because it was also pegging the CPU and making my other services be unresponsive.
Correct, if you share the same VM, machine with other services, you can limit the CPU usage of the container to avoid one takes all the CPU processing power from others