Page tree
Skip to end of metadata
Go to start of metadata

Regular weekly maintenance windows

In order to more quickly and regularly bring new Atlassian product features and fixes to you, we update your instances in weekly maintenance windows at the times listed below.

Your regionWindow

APAC (Asia Pacific)

every Monday 23:00 - Tuesday 01:00, Sydney time

EMEA (Europe, Middle East and Africa, including India)

every Monday 02:00 - 05:00, Amsterdam time
AMER (America, including U.S.)every Sunday 17:00 - 20:00, San Francisco time

Atlassian Customer Agreement

Icon

Click here to view the new Atlassian Customer Agreement that goes into effect on April 28, 2014.

Getting notified about upcoming upgrades and news

We use the Atlassian OnDemand blog to announce upcoming upgrades and the completion of them. We try to publish information about upcoming upgrades a few days prior to the weekly maintenance so that you can prepare for them if needed.

If you'd like to get notified about future Atlassian OnDemand upgrades and news,  please subscribe to our blog.

 How to subscribe...
  1. Sign up for a username on this site in 15 seconds, and log in.
  2. At the top right of any blog post, choose Watch and then select Watch for new blog posts in this space. You will then get email notifications when new blog posts are published.
  3. If you'd like to participate in the discussion and be notified of new comments on a particular blog post, choose Watch > Watch blog post on that post. 

 

Communication on maintenance windows

  • Regular weekly maintenance windows: We will not send maintenance notifications unless we need to deviate from the weekly schedule.
  • Other types of maintenance: For other required maintenance such as urgent fixes, we will notify you at least 12 hours in advance, by sending an email message to all technical contacts listed under your account at my.atlassian.com.

Related topics

Where do I set the region used for maintenance windows?

20 Comments

  1. Does anyone know if the weekly maintenance causes an outage every week?  It seems like we are still able to access our environment during the window.

    Regards,

    Pan

    1. Hi Pan,

      If there is maintenance that requires restarting your services, this will cause an outage.  The window is there if we need it, but there may not always be an outage.

      Thanks,

      -dave

      1. Thanks for the quick reply, Dave!

        We are trying to figure this out because we are using OnDemend US Region, but we have offices in Sydney & Tokyo where the maintenance window is in the beginning of the work week.  Furthermore, we have a 24x7 support team that is dependent on our documentation written in our Confluence pages.  I know that we post what is getting upgraded for the weekly maintenance in https://confluence.atlassian.com/display/AOD/What%27s+New.  But is there a way to tell if the particular week's maintenance will have an outage or not?  It would be good to know so beforehand, so we can inform our Sydney & Tokyo's co-workers.

        1. Hi Pan,

          Basically, you should expect the system to be unavailable during this time each week, and communicate that to all of your teams.  Some updates can simply be pushed without interruption and some will will require restarting the system, which is generally less than 30 minutes.  To be safe and set expectations appropriately, there is a 3-hour window each week when we do this.

          Thanks,

          -dave

  2. Atlassian really needs a third time window that is better for the Americas. I see we're lumped together with EMEA, and this creates a problem when my company is getting ready for the week and Sunday evening they have no access to JIRA. 

  3. Sorry, but I couldn't find Brazil. Is it "U.S."? If it is the case, wouldn't it be better America instead?

  4. Anonymous

    It says APAC maintenance is every Monday 23:00 - Tuesday 01:00, Sydney time but its now 12.28pm on a monday and the site has been down for us since weekend and still unaccessible.

    1. If your OnDemand instance has not been accessible all weekend you should contact support for assistance.  

  5. These maintenance windows are absurd. If you are servicing an area, you should respect the times that they are utilizing the system. I am pretty sure the majority of your customers are IT/PM teams. All of us understand the importance of maintenance; however, it's a bit absurd to do this within business hours for the time zone. Since maintenance usually simply means a backup of the server/databases, we all know that this is in most part automated. Why not automate them during an hour where it's not business hours for your customers?

    And before I get one of the generic responses, Atlassian should truly stop with replies that say this is just the way we do it. I have several outstanding tickets that haven't been handled in years. That alone was enough to upset the leaders in our company. It's starting to reach a point where I'm going to be tasked with developing our own system since the product is becoming less of a solution and more of a problem. 

    1. Hi Mack,  We try our best to make sure the maintenance windows fall outside standard business hours in each timezone, as listed above. Is your OnDemand instance perhaps used across multiple timezones, or was it set up in one region but now used in another? (for example the original billing address was in the US, but you are accessing in Australia?).  This might be why you are seeing outages in your business hours. 

      Please raise a support request in the "OnDemand Support" project at https://support.atlassian.com and ask them to check which region your instance is associated with, and whether it can be changed to a more suitable region.

  6. I received an advisory stating that maintenance will be done this Saturday outside our normal maintenance window. The notice says: 09:11 - 11:00. Is this based on a 24-hr clock? If not, how can I tell if it is AM or PM?

    1. Sorry for the confusion Bob, it's based on 24 hour time. Right below the times we also linked a page that tells you what time it will be in your local timezone.

  7. Thanks for the quick response, Matthew. I tried the link and thought that the case, but wasn't 100% sure. I appreciate the clarification.

  8. Hi,

    JIRA On demand will be down for 3h from 1pm UK time this afternoon. I realise you can't really win, but the timing could be a lot better for customers in and around our timezone. It must be important.

    James

    1. If you're being upgraded during your business hours then your instance may have it's timezone set incorrectly (i.e. your instance is set to be in Sydney instead of the UK). Please log a ticket at https://support.atlassian.com if you want to investigate further.

  9. I'm surprised you guys consider Sunday night to be a good maintenance window? This is the time where a lot of my peers get things together for the beginning of the next week. Maybe you could consider Friday or Saturday night? Or even Sunday morning?

    1. This is exactly what I was talking about Chad. JIRA is a product that is in most part used by programmers. It's crazy that they have these backups performed during hours that are actually working hours for a lot of programmers. For instance, the following are the times that "AMER" clients are unable to use the system.

      • Customers in Pacific Time: 5:00pm - 8:00pm
      • Customers in Mountain Time: 6:00pm - 9:00pm
      • Customers in Central Time: 7:00pm - 10:00pm
      • Customers in Eastern Time: 8:00pm - 11:00pm

      Why would they not move it to something more understandable like:

      • Pacific Time: 12:00am - 3:00am
      • Mountain Time: 1:00am - 4:00am
      • Central Time: 2:00am - 5:00am
      • Eastern Time: 3:00am - 6:00am

      Additionally, why aren't they looking into why it's taking 3 hours to backup the AMER (or any to be honest) server. In any case, they at least provide a solution until I have the time to develop something else I guess.

      1. The US and EMEA upgrades are currently scheduled at a time that we hope will inconvenience the least number of customers. If you feel that the window is inconvenient for your timezone then please vote for  AOD-7050 - Perform US upgrades overnight rather than in the afternoon Open - untriaged .

        That being said though, your instance will generally not be down for the entire duration of the maintenance window, it will just be down for some portion of it.

        Additionally, why aren't they looking into why it's taking 3 hours to backup the AMER (or any to be honest) server. In any case, they at least provide a solution until I have the time to develop something else I guess.

        During the maintenance window we're installing new versions of our software to resolve bugs and release new features. Backups are performed seamlessly every day as per the article 'About Your Data'. Backing up your instance will not cause an outage.

         

        1. In all honesty, it's great that you decided to create an issue; however, I have a couple AOD issues out there I've flowed out, justified and provided lengthy detail on, which are years old. There has been no movement on them and in all honesty those of a much higher priority. The main reason for that is they allow those handling issues to actually see the issues dispositioned differently when the issue is commented on.

          So back on topic. You're saying that every Sunday AOD is updated to resolve bugs and install new features? If so, do you have a listing of all these new features and bugs that are being fixed. There must be a ton of them as ours has gone down like clockwork for several months now. In regards to the length of downtime, it is actually down for almost the entire 3 hour period. If it were down for only 15 minutes, I would hardly be making a stink about it.

          1. In all honesty, it's great that you decided to create an issue; however, I have a couple AOD issues out there I've flowed out, justified and provided lengthy detail on, which are years old. There has been no movement on them and in all honesty those of a much higher priority. The main reason for that is they allow those handling issues to actually see the issues dispositioned differently when the issue is commented on.

            I understand your frustration, while we try to implement as many highly voted issues and fix as many bugs as possible we do need to perform some prioritisation, which we do by following our Implementation of New Features Policy and Atlassian Bug Fixing Policy.

            So back on topic. You're saying that every Sunday AOD is updated to resolve bugs and install new features? If so, do you have a listing of all these new features and bugs that are being fixed. 

            You can find a summary of all releases as a child of What's New. For future releases you should subscribe to the Atlassian OnDemand Blog. If you're interested in more granular information you can have a read of the Production Release Notes for JIRA, the point releases include information on specific bugs that are fixed as a part of each release. These don't map up directly to OnDemand releases, but OnDemand will generally receive fixes before the OnPremise version of the product.

            There must be a ton of them as ours has gone down like clockwork for several months now

            If your instance is going down regularly please log a support ticket, as it certainly shouldn't be doing that.

            In regards to the length of downtime, it is actually down for almost the entire 3 hour period. If it were down for only 15 minutes, I would hardly be making a stink about it.

            Your instance also shouldn't be down for almost the entire 3 hour maintenance window. I'll keep my eye on it next upgrade window to see if it's staying down, and why that would be the case.