Bug Tracking Dilemma

March 11, 2009 3:15 PM

Bug Trackers: HiveMind, I need your opinions on the various bug tracking applications out there. I'm in something of a Goldilocks situation. Whole story inside.

I'm the internal tools director for a software development firm. Prior to this job, I'm just 4 months in, I've had no previous experience with bug trackers. My corner of the enterprise uses JIRA. It's open source, cheap, flexible, accepted by my users, and web-based. It's also a bit weak, can only handle 200K bugs per instance (we have 400k currently and add another 10k per month), and a bit fragile.

The rest of the enterprise uses DevTrack. It does not have a limit on the number of bugs, but is not designed to integrate multiple projects into searches, profiles and reporting. This is kind of a deal-breaker for my organization.

So the question is, what's a good bug tracker that can handle a million plus bugs across hundreds of separate projects, and can slice and dice data across any given data field? Any good experiences with Bugzilla? Bis4Bug? Anything else?

posted by bluejayway to technology (14 comments total)

I've always used Trac. It'll store its database in a real database (I use PostgreSQL) so limits are basically moot (IIRC PostgreSQL has a hard limit around 4 trillion rows per table). I like that it's simple, has customizable workflow, a good plugin API and lots of available open source plugins, and integrates bug tracking, a wiki, and source control. I'm happy to give more details if you like; just ask.

OTOH, I've hated every Bugzilla instance I've ever had to touch either as a user or as an admin.
posted by jacobian at 3:21 PM on March 11

I use and like Trac, now that it's running. The initial setup was more painful than I expected, though (ran into complexities either with MySQL or SVN, I forget which).
posted by zippy at 3:24 PM on March 11

I have no direct experience with it, but FogBugz (FogCreek.com) looks worth considering.
posted by TruncatedTiller at 3:25 PM on March 11

There's a pretty good discussion of this over on Stack Overflow.

Personally, I'm partial to either trac or bugzilla, but a lot depends on your institution. If you guys are tracking that many bugs on a single instance, I'd definitely look at some of the commercial products.
posted by jenkinsEar at 3:32 PM on March 11

Jira isn't open source is it?

We use jira at work and it's very impressive. I've used Bugzilla which is good, but not as good as jira, it wasn't as easily customisable.
posted by sien at 3:59 PM on March 11

I've used Mantis for over 6 years now. For us, much depended on how easy we could customize, and now we use Mantis for almost everything. I'm sure it can handle millions of bugs, since it is just a matter of throwing a good server in. The only thing I dislike is it's search engine, although we're using a pretty old verision now, and I'm sure it has been enhanced since then.
posted by theKik at 5:13 PM on March 11

Second Mantis. Very easy to set up on even the most bare-bones system, easy enough to modify to suit your own needs. We've been using it at my office since I came on board and liked it so much we're also using it for issue tracking and project planning / prioritizing. It's basically idiot-proof, and doesn't have the HOLYSHITOPTIONSSELECTSMOTHERFUCKININPUTSPRIORITIESSEVERITIES look of more "professional" bug trackers like JIRA / BugZilla.
posted by Civil_Disobedient at 5:48 PM on March 11

I think JIRA is Free As In Beer At A Bar.

Like Java, JIRA is cumbersome and flawed, but it does the job. It has been a long time since I've used Mantis, but it seemed to do the job at the time.
posted by b1tr0t at 6:42 PM on March 11

Yeah, Mantis makes workflow a lot better. It color-codes issues, makes them far easier to track, and isn't all cluttered and clusterfucky like BugZilla. I absolutely hate BZ and its interface--it's a complete nightmare, really.
posted by disillusioned at 6:44 PM on March 11

I haven't found JIRA to be fragile. Where's the 200k per instance number come from? Are you using the embedded database engine (HSQLDB) or are you using MySQL or another "real" database engine behind it?

We ended up on JIRA because it's the only one that works well for us in all roles: programmers who need subversion integration, server support people with big long-running internal projects and lots of task dependencies, and desktop support people with lots of user-interaction.
posted by SpecialK at 7:18 PM on March 11

We use bugzilla and it's pretty ugly, but does the job.
posted by mattoxic at 9:24 PM on March 11

Where's the 200k per instance number come from? Are you using the embedded database engine (HSQLDB) or are you using MySQL or another "real" database engine behind it?


That figure comes directly from Atlassian. We are using an Oracle10g database behind it. All of our troubles have come from the application side- the db is holding up fine.

Thanks all for the opinions. I will definitely check out Mantis and Trac.
posted by bluejayway at 2:00 PM on March 12

We ended up on JIRA because it's the only one that works well for us in all roles: programmers who need subversion integration[...]

Just FYI and for posterity: you can, technically, get Mantis to work with Subversion. It is not terribly straight-forward, but definitely possible.
posted by Civil_Disobedient at 6:58 PM on March 15

FWIW, Trac works pretty nicely with subversion.
posted by zippy at 6:25 PM on March 24