jira slow or temporarily unavailable

Give me an ssh key in pm and you can :stuck_out_tongue: – please keep the private key somewhat secure :slight_smile:

2 Likes

I realised we may also have to host the database, so we’re looking at more like $40/mo.

Off the top of my head: branding, single sign-on, existing content, and potentially cost.

In the Atlassian products I’ve managed, I’m a big fan of limiting add-ons to only the very few that are robust and well-supported (i.e., regularly released ahead of Atlassian releases) like Balsamiq, Gliffy, and Zephyr. Avoiding other add-ons and never touching SQL under the hood help ensure upgrading is smooth.

Here’s a list of add-ons running on JIRA. I’ve removed some I know are supplied by Atlassian out of the box; some of these may fall into that category as well:

  • Announcer for JIRA
  • Application Links Link Rendering Plugin
  • Balsamiq Mockups
  • Bitbucket Importer Plugin
  • Gliffy JIRA Plugin
  • JIRA Charting Plugin
  • JIRA iCalendar Plugin
  • JIRA Importers Plugin (JIM)
  • JIRA Misc Workflow Extensions
  • JIRA Toolkit Plugin
  • JIRA Workflow Designer Plugin
  • Links Hierarchy Reports
  • Redmine Importers Plugin for JIM
  • ServiceRocket JIRA Utilities
  • SuggestiMate for JIRA
  • SuggestiMate for JIRA
  • Wallboard Plugin
  • Zephyr for JIRA

Not too bad.

Per this page, a Micro won’t do. They recommend a M3 or C3. For our community and size, I’d assume we’d need something like a c3.xlarge. From the pricing page, it looks like in USD a c3.xlarge instance ranges from $688.67/year ($2066 paid up front for 3 yr term) up to $1844.64/year ($0.21/hour on demand on a leap year).

2 Likes

I’m not sure we need that much (although I’m sure it would be great). The c3.xlarge has 7.5GB of RAM and at the moment we’re somewhat managing with just 1GB.

That cost seems reasonable for a product serving the entire community. Does that include us being able to hand off some of the tasks we are currently doing on the infra team to support JIRA or would those still remain with the community?

When you do any on-boarding to infrastructure, please take copious notes. We can use those notes to develop some on-boarding documentation. To facilitate scaling up the support team, I think we need to move away from an all-or-none model – i.e., we should be able to bring someone to support a specific application or service. To help with this, I’ve created a hierarchy of folders in the shared Google Drive Infrastructure folder both to organize documentation and also allow us to share contents at an appropriate level.

It looks like Atlassian recommends nothing less than 3.75 GB (at least when directing people to AWS).

1 Like

We’d still maintain that server.

That is understandable and I’m not asking for the keys to the kingdom but, help in anyway I can. I don’t expect to be handed everything either. I’ll help in anyway I can suggestions, hints research does not matter.

2 Likes

I do appreciate your help – jus can’t hand anyone access to production systems. Already had one case where some people did something to one of our staging servers, so i’m not even more vigilant with who I grant access.

no problem. I’ll help the way I can by researching what Darious asked before. Then we can go from there.

2 Likes

Definitely an out of memory:

OutOfMemory due to Large Mailbox Processing https://confluence.atlassian.com/jirakb/outofmemory-due-to-large-mailbox-processing-216433923.html?utm_medium=logScan&utm_source=STP

OutOfMemory due to Mail Loop OutOfMemory due to Mail Loop | Jira | Atlassian Documentation? utm_medium=logScan&utm_source=STP

OutOfMemory Errors Due to Third-Party JIRA Plugins https://confluence.atlassian.com/jirakb/outofmemory-errors-due-to-third-party-jira-plugins-216433918.html?utm_medium=logScan&utm_source=STP

OutOfMemory due to Remote API https://confluence.atlassian.com/jirakb/outofmemory-due-to-remote-api-216433936.html?utm_medium=logScan&utm_source=STP

OutOfMemoryError caused by a large number of Issue Security Levels https://jira.atlassian.com/browse/JRA-19622?utm_source=STP&utm_medium=logScan

1 Like

yeah there are ways to fix those. Specially the mail box part. At my job last year we had those mail loop and it took everything down.

There are 66k users in OpenMRS JIRA. It’s a huge number, can it affect performance?

1 Like

it does to a point. I don’t know how to explain that part but, if at least around 1k-3k are log-in at then it does impact performance with 1G in JVM.

1 Like

also what DB is used? maybe separating the db to another host might help? just trowing suggestions since I’m not sure what is the budget for any of the solutions or how the infrastructure connects together.

1 Like

DB on another host just opens up a can of worms security wise.

We use mysql.

I’m not sure what you guys do with this add-on’s but, I’ll give my two cents:

Announcer for JIRA - everything done in this plug-in can be done in confluence. The only good part is the banner that you can use to put out messages or other things.

Application Links Link Rendering Plugin - This plug-in is no longer supported https://marketplace.atlassian.com/archive/com.atlassian.linkers.applinks-linkers and I believe this was added into Jira.

Balsamiq Mockups - How much is this used? this can be removed depending on the usage and how helpful it is.

Bitbucket Importer Plugin - This can be removed if you are not expecting any imports to be done soon. Not sure if this part of Jira or just an extra thing. We use Github/Gitlab and have used bitbucket in a while.

JIRA Charting Plugin - this is helpful to have more metrics and other things.

JIRA iCalendar Plugin - this plugin is needed and can break certain things if removed.

JIRA Importers Plugin (JIM) - This one is part of Jira.

maybe we can use this plugin to help with the email problem?

I know theres a Jira plugin that helps with email storms and the way emails are generated but, I can’t remember right now. Anyways this are just suggestions so is up to you guys on what to do.

2 Likes

@chagara, you don’t happen to have a cv do you? I’m beginning to see the value you could add :)or just say what you’ve done and such in a pm?

1 Like

What if we load balanced across servers? Never set this up but willing to learn.

Atlassian did come out with an datacenter version that supports out of the box load balancing and better high availability. https://www.atlassian.com/enterprise/data-center Not sure the cost of it for this situation. Besides this I think having two different regular Jira instances might be a problem with making sure the data its correct and both have the latest( tickets and such). As far as I know Jira can only access one DB. Might be a little more complicated but, it can be research if we do not have options.

Also it would help if Jira gets updated as a log of bugs have been fixed since v6.2.1 but, that would require testing, integration, planning, backup etc etc.