#1996 - [GH-ISSUE #99] Meshcentral - 502 Bad Gateway nginx (reboot fixes for a period of time) - starred/tacticalrmm

kerem commented

2026-03-14 02:00:46 +03:00

Owner

Originally created by @rtwright68 on GitHub (Sep 10, 2020).
Original GitHub issue: https://github.com/amidaware/tacticalrmm/issues/99

Within the last couple of updates, I run into 502 Bad Gateway - nginx while trying to take control of an agent via Meshcentral.
Rebooting the server fixes it for a period of time. I rebooted this morning and now this afternoon when going back into it I am seeing the same error.
Not sure what to take a look at?

Originally created by @rtwright68 on GitHub (Sep 10, 2020). Original GitHub issue: https://github.com/amidaware/tacticalrmm/issues/99 Within the last couple of updates, I run into 502 Bad Gateway - nginx while trying to take control of an agent via Meshcentral. Rebooting the server fixes it for a period of time. I rebooted this morning and now this afternoon when going back into it I am seeing the same error. Not sure what to take a look at?

kerem closed this issue

2026-03-14 02:00:51 +03:00

kerem commented

2026-03-14 02:00:57 +03:00

Author

Owner

@dinger1986 commented on GitHub (Sep 10, 2020):

I’m also getting that.

Out of interest how much free memory do you have on your machine? Or how much did you start with?

I have 2gb and it’s only got about 10% free after being in for 2 days....

Daniel Lamb

Technical Director

[cid:logo_2094e808-adcc-455a-a5fd-54265dc137fa.jpg]http://www.flonix.co.uk/

Flonix Limited
Willowview
Redgorton
Perth
PH1 3EL

Phone: 01738 500400

http://www.flonix.co.uk

Company Reg: SC341802

[cid:fb_11c7c55d-b85e-4b3e-b8d8-daabbed1905e.png]https://www.facebook.com/Flonixltd/

From: rtwright68 notifications@github.com
Sent: Thursday, September 10, 2020 7:20:06 PM
To: wh1te909/tacticalrmm tacticalrmm@noreply.github.com
Cc: Subscribed subscribed@noreply.github.com
Subject: [wh1te909/tacticalrmm] Meshcentral - 502 Bad Gateway nginx (reboot fixes for a period of time) (#99)

Within the last couple of updates, I run into 502 Bad Gateway - nginx while trying to take control of an agent via Meshcentral.
Rebooting the server fixes it for a period of time. I rebooted this morning and now this afternoon when going back into it I am seeing the same error.
Not sure what to take a look at?

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHubhttps://github.com/wh1te909/tacticalrmm/issues/99, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ABXIVH2VAC4XZ7TDUTESXYDSFEKFNANCNFSM4RFSS5WA.

@dinger1986 commented on GitHub (Sep 10, 2020): I’m also getting that. Out of interest how much free memory do you have on your machine? Or how much did you start with? I have 2gb and it’s only got about 10% free after being in for 2 days.... Daniel Lamb Technical Director [cid:logo_2094e808-adcc-455a-a5fd-54265dc137fa.jpg]<http://www.flonix.co.uk/> Flonix Limited Willowview Redgorton Perth PH1 3EL Phone: 01738 500400 http://www.flonix.co.uk Company Reg: SC341802 [cid:fb_11c7c55d-b85e-4b3e-b8d8-daabbed1905e.png]<https://www.facebook.com/Flonixltd/> ________________________________ From: rtwright68 <notifications@github.com> Sent: Thursday, September 10, 2020 7:20:06 PM To: wh1te909/tacticalrmm <tacticalrmm@noreply.github.com> Cc: Subscribed <subscribed@noreply.github.com> Subject: [wh1te909/tacticalrmm] Meshcentral - 502 Bad Gateway nginx (reboot fixes for a period of time) (#99) Within the last couple of updates, I run into 502 Bad Gateway - nginx while trying to take control of an agent via Meshcentral. Rebooting the server fixes it for a period of time. I rebooted this morning and now this afternoon when going back into it I am seeing the same error. Not sure what to take a look at? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub<https://github.com/wh1te909/tacticalrmm/issues/99>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ABXIVH2VAC4XZ7TDUTESXYDSFEKFNANCNFSM4RFSS5WA>.

kerem commented

2026-03-14 02:01:02 +03:00

Author

Owner

@rtwright68 commented on GitHub (Sep 10, 2020):

I also have a 2GB droplet via DigitalOcean. I have 639,928 free currently (rebooted around 5AM EST this morning). Very small set of agents (35 currently testing)

@rtwright68 commented on GitHub (Sep 10, 2020): I also have a 2GB droplet via DigitalOcean. I have 639,928 free currently (rebooted around 5AM EST this morning). Very small set of agents (35 currently testing)

kerem commented

2026-03-14 02:01:07 +03:00

Author

Owner

@dinger1986 commented on GitHub (Sep 10, 2020):

I only have 10 agents at the moment for testing but planned on rolling out next week, I am planning on increasing to 4GB ram and see how it goes.

Daniel

Daniel Lamb

Technical Director

[cid:logo_2094e808-adcc-455a-a5fd-54265dc137fa.jpg]http://www.flonix.co.uk/

Flonix Limited
Willowview
Redgorton
Perth
PH1 3EL

Phone: 01738 500400

http://www.flonix.co.uk

Company Reg: SC341802

[cid:fb_11c7c55d-b85e-4b3e-b8d8-daabbed1905e.png]https://www.facebook.com/Flonixltd/

From: rtwright68 notifications@github.com
Sent: 10 September 2020 19:34
To: wh1te909/tacticalrmm tacticalrmm@noreply.github.com
Cc: Daniel Lamb | Flonix Ltd daniel.lamb@flonix.co.uk; Comment comment@noreply.github.com
Subject: Re: [wh1te909/tacticalrmm] Meshcentral - 502 Bad Gateway nginx (reboot fixes for a period of time) (#99)

I also have a 2GB droplet via DigitalOcean. I have 639,928 free currently (rebooted around 5AM EST this morning). Very small set of agents (35 currently testing)

—
You are receiving this because you commented.
Reply to this email directly, view it on GitHubhttps://github.com/wh1te909/tacticalrmm/issues/99#issuecomment-690606632, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ABXIVHY4732QLFVFDWH24JDSFEL3DANCNFSM4RFSS5WA.

@dinger1986 commented on GitHub (Sep 10, 2020): I only have 10 agents at the moment for testing but planned on rolling out next week, I am planning on increasing to 4GB ram and see how it goes. Daniel Daniel Lamb Technical Director [cid:logo_2094e808-adcc-455a-a5fd-54265dc137fa.jpg]<http://www.flonix.co.uk/> Flonix Limited Willowview Redgorton Perth PH1 3EL Phone: 01738 500400 http://www.flonix.co.uk Company Reg: SC341802 [cid:fb_11c7c55d-b85e-4b3e-b8d8-daabbed1905e.png]<https://www.facebook.com/Flonixltd/> ________________________________ From: rtwright68 <notifications@github.com> Sent: 10 September 2020 19:34 To: wh1te909/tacticalrmm <tacticalrmm@noreply.github.com> Cc: Daniel Lamb | Flonix Ltd <daniel.lamb@flonix.co.uk>; Comment <comment@noreply.github.com> Subject: Re: [wh1te909/tacticalrmm] Meshcentral - 502 Bad Gateway nginx (reboot fixes for a period of time) (#99) I also have a 2GB droplet via DigitalOcean. I have 639,928 free currently (rebooted around 5AM EST this morning). Very small set of agents (35 currently testing) — You are receiving this because you commented. Reply to this email directly, view it on GitHub<https://github.com/wh1te909/tacticalrmm/issues/99#issuecomment-690606632>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ABXIVHY4732QLFVFDWH24JDSFEL3DANCNFSM4RFSS5WA>.

kerem commented

2026-03-14 02:01:13 +03:00

Author

Owner

@rtwright68 commented on GitHub (Sep 10, 2020):

Good idea. I will do the same. Next step up for DigitalOcean is 2vCPU and 4GB

@rtwright68 commented on GitHub (Sep 10, 2020): Good idea. I will do the same. Next step up for DigitalOcean is 2vCPU and 4GB

kerem commented

2026-03-14 02:01:18 +03:00

Author

Owner

@Omnicef commented on GitHub (Sep 10, 2020):

I had the same issue at 2GB. Bumping up to 4GB solved it for me.

@Omnicef commented on GitHub (Sep 10, 2020): I had the same issue at 2GB. Bumping up to 4GB solved it for me.

kerem commented

2026-03-14 02:01:23 +03:00

Author

Owner

@dinger1986 commented on GitHub (Sep 10, 2020):

Easy to do on vultr, just change plan so going for it now

Daniel Lamb

Technical Director

[cid:logo_2094e808-adcc-455a-a5fd-54265dc137fa.jpg]http://www.flonix.co.uk/

Flonix Limited
Willowview
Redgorton
Perth
PH1 3EL

Phone: 01738 500400

http://www.flonix.co.uk

Company Reg: SC341802

[cid:fb_11c7c55d-b85e-4b3e-b8d8-daabbed1905e.png]https://www.facebook.com/Flonixltd/

From: rtwright68 notifications@github.com
Sent: 10 September 2020 19:36
To: wh1te909/tacticalrmm tacticalrmm@noreply.github.com
Cc: Daniel Lamb | Flonix Ltd daniel.lamb@flonix.co.uk; Comment comment@noreply.github.com
Subject: Re: [wh1te909/tacticalrmm] Meshcentral - 502 Bad Gateway nginx (reboot fixes for a period of time) (#99)

Good idea. I will do the same. Next step up for DigitalOcean is 2vCPU and 4GB

—
You are receiving this because you commented.
Reply to this email directly, view it on GitHubhttps://github.com/wh1te909/tacticalrmm/issues/99#issuecomment-690608326, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ABXIVH3I5BOVEJOYAMLN7YLSFEMD3ANCNFSM4RFSS5WA.

@dinger1986 commented on GitHub (Sep 10, 2020): Easy to do on vultr, just change plan so going for it now Daniel Lamb Technical Director [cid:logo_2094e808-adcc-455a-a5fd-54265dc137fa.jpg]<http://www.flonix.co.uk/> Flonix Limited Willowview Redgorton Perth PH1 3EL Phone: 01738 500400 http://www.flonix.co.uk Company Reg: SC341802 [cid:fb_11c7c55d-b85e-4b3e-b8d8-daabbed1905e.png]<https://www.facebook.com/Flonixltd/> ________________________________ From: rtwright68 <notifications@github.com> Sent: 10 September 2020 19:36 To: wh1te909/tacticalrmm <tacticalrmm@noreply.github.com> Cc: Daniel Lamb | Flonix Ltd <daniel.lamb@flonix.co.uk>; Comment <comment@noreply.github.com> Subject: Re: [wh1te909/tacticalrmm] Meshcentral - 502 Bad Gateway nginx (reboot fixes for a period of time) (#99) Good idea. I will do the same. Next step up for DigitalOcean is 2vCPU and 4GB — You are receiving this because you commented. Reply to this email directly, view it on GitHub<https://github.com/wh1te909/tacticalrmm/issues/99#issuecomment-690608326>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ABXIVH3I5BOVEJOYAMLN7YLSFEMD3ANCNFSM4RFSS5WA>.

kerem commented

2026-03-14 02:01:28 +03:00

Author

Owner

@rtwright68 commented on GitHub (Sep 10, 2020):

Yup, have mine changed on DigitalOcean already. Will keep an eye on memory usage.

@rtwright68 commented on GitHub (Sep 10, 2020): Yup, have mine changed on DigitalOcean already. Will keep an eye on memory usage.

kerem commented

2026-03-14 02:01:33 +03:00

Author

Owner

@dinger1986 commented on GitHub (Sep 10, 2020):

I have upgraded as well so see if it behaves, looking good at the moment as didnt reboot and mesh is now online again.

Daniel

Daniel Lamb

Technical Director

[cid:logo_2094e808-adcc-455a-a5fd-54265dc137fa.jpg]http://www.flonix.co.uk/

Flonix Limited
Willowview
Redgorton
Perth
PH1 3EL

Phone: 01738 500400

http://www.flonix.co.uk

Company Reg: SC341802

[cid:fb_11c7c55d-b85e-4b3e-b8d8-daabbed1905e.png]https://www.facebook.com/Flonixltd/

From: rtwright68 notifications@github.com
Sent: 10 September 2020 19:51
To: wh1te909/tacticalrmm tacticalrmm@noreply.github.com
Cc: Daniel Lamb | Flonix Ltd daniel.lamb@flonix.co.uk; Comment comment@noreply.github.com
Subject: Re: [wh1te909/tacticalrmm] Meshcentral - 502 Bad Gateway nginx (reboot fixes for a period of time) (#99)

Yup, have mine changed on DigitalOcean already. Will keep an eye on memory usage.

—
You are receiving this because you commented.
Reply to this email directly, view it on GitHubhttps://github.com/wh1te909/tacticalrmm/issues/99#issuecomment-690618543, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ABXIVH5QJD4ZVWNBV3VP4VTSFEN4TANCNFSM4RFSS5WA.

@dinger1986 commented on GitHub (Sep 10, 2020): I have upgraded as well so see if it behaves, looking good at the moment as didnt reboot and mesh is now online again. Daniel Daniel Lamb Technical Director [cid:logo_2094e808-adcc-455a-a5fd-54265dc137fa.jpg]<http://www.flonix.co.uk/> Flonix Limited Willowview Redgorton Perth PH1 3EL Phone: 01738 500400 http://www.flonix.co.uk Company Reg: SC341802 [cid:fb_11c7c55d-b85e-4b3e-b8d8-daabbed1905e.png]<https://www.facebook.com/Flonixltd/> ________________________________ From: rtwright68 <notifications@github.com> Sent: 10 September 2020 19:51 To: wh1te909/tacticalrmm <tacticalrmm@noreply.github.com> Cc: Daniel Lamb | Flonix Ltd <daniel.lamb@flonix.co.uk>; Comment <comment@noreply.github.com> Subject: Re: [wh1te909/tacticalrmm] Meshcentral - 502 Bad Gateway nginx (reboot fixes for a period of time) (#99) Yup, have mine changed on DigitalOcean already. Will keep an eye on memory usage. — You are receiving this because you commented. Reply to this email directly, view it on GitHub<https://github.com/wh1te909/tacticalrmm/issues/99#issuecomment-690618543>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ABXIVH5QJD4ZVWNBV3VP4VTSFEN4TANCNFSM4RFSS5WA>.

kerem commented

2026-03-14 02:01:38 +03:00

Author

Owner

@wh1te909 commented on GitHub (Sep 10, 2020):

I upgraded the minimum requirements a week or so ago on the readme to 4GB thanks to @Omnicef's suggestion, I have a feeling that salt leaks memory and needs to be restarted every few days, and possibly other services too. If you don't want to restart the server here are all the services that you can restart to free up any memory leaks: salt-master salt-api nginx rmm celery celerybeat celery-winupdate meshcentral Nginx needs to be running before meshcentral otherwise order should not matter

And yea the more agents you add the more memory you'll need. With my 600 agents i am averaging around 5GB usage, however I have also increased the amount of uwsgi workers and salt-master and salt-api threads to handle the load of that many agents, which increases memory usage. You can play around with the number of uwsgi workers in /rmm/api/tacticalrmm/app.ini (processes and threads) and for salt, in /etc/salt/master.d/rmm-salt.conf

here is my rmm-salt.conf (notice the 2 extra fields: worker_threads for salt-master, and thread_pool for salt-api)
default if none are specified are 5 for worker_threads and 100 for thread_pool, but these should not be lowered only increased.

if you change any of these numbers in app.ini or rmm-salt.conf then make sure to restart all the services above to load the new configs

timeout: 20
worker_threads: 20
gather_job_timeout: 25
max_event_size: 30485760
external_auth:
  pam:
    saltapi:
      - .*
      - '@runner'
      - '@wheel'
      - '@jobs'

rest_cherrypy:
  port: 8123
  disable_ssl: True
  max_request_body_size: 30485760
  thread_pool: 300

@wh1te909 commented on GitHub (Sep 10, 2020): I upgraded the minimum requirements a week or so ago on the readme to 4GB thanks to @Omnicef's suggestion, I have a feeling that salt leaks memory and needs to be restarted every few days, and possibly other services too. If you don't want to restart the server here are all the services that you can restart to free up any memory leaks: ```salt-master salt-api nginx rmm celery celerybeat celery-winupdate meshcentral``` Nginx needs to be running before meshcentral otherwise order should not matter And yea the more agents you add the more memory you'll need. With my 600 agents i am averaging around 5GB usage, however I have also increased the amount of uwsgi workers and salt-master and salt-api threads to handle the load of that many agents, which increases memory usage. You can play around with the number of uwsgi workers in ```/rmm/api/tacticalrmm/app.ini``` (```processes``` and ```threads```) and for salt, in ```/etc/salt/master.d/rmm-salt.conf``` here is my rmm-salt.conf (notice the 2 extra fields: worker_threads for salt-master, and thread_pool for salt-api) default if none are specified are 5 for ```worker_threads``` and 100 for ```thread_pool```, but these should not be lowered only increased. if you change any of these numbers in ```app.ini``` or ```rmm-salt.conf``` then make sure to restart all the services above to load the new configs ``` timeout: 20 worker_threads: 20 gather_job_timeout: 25 max_event_size: 30485760 external_auth: pam: saltapi: - .* - '@runner' - '@wheel' - '@jobs' rest_cherrypy: port: 8123 disable_ssl: True max_request_body_size: 30485760 thread_pool: 300 ```

kerem commented

2026-03-14 02:01:43 +03:00

Author

Owner

@wh1te909 commented on GitHub (Oct 25, 2020):

timeout: 20
worker_threads: 15
gather_job_timeout: 25
max_event_size: 30485760
external_auth:
  pam:
    saltapi:
      - .*
      - '@runner'
      - '@wheel'
      - '@jobs'
rest_cherrypy:
  port: 8123
  disable_ssl: True
  max_request_body_size: 30485760
  thread_pool: 300
  socket_queue_size: 100

I've updated the defaults in the install script to the above, just found out about the last line socket_queue_size so you guys can add that too should help with the salt api load.

@wh1te909 commented on GitHub (Oct 25, 2020): ``` timeout: 20 worker_threads: 15 gather_job_timeout: 25 max_event_size: 30485760 external_auth: pam: saltapi: - .* - '@runner' - '@wheel' - '@jobs' rest_cherrypy: port: 8123 disable_ssl: True max_request_body_size: 30485760 thread_pool: 300 socket_queue_size: 100 ``` I've updated the defaults in the install script to the above, just found out about the last line ```socket_queue_size``` so you guys can add that too should help with the salt api load.

kerem referenced this issue

2026-03-14 06:50:11 +03:00

[GH-ISSUE #1996] Enhance Script Manager UI #3195

Rows
Columns

[GH-ISSUE #99] Meshcentral - 502 Bad Gateway nginx (reboot fixes for a period of time) #1996