How to keep performance humming during this defining era and what to look out for.
Long coveted by many employees, remote work is something that has been picking up steam with organizations for years. Recently, we have seen a seismic shift as companies look to attract and manage globally dispersed talent supported by distributed work. This challenge has inspired innovation and changed the way we all communicate and collaborate online and, as such, organizations have been tasked with mitigating the negative side effects of this shift without missing a beat.
Enterprises, in general, are acutely aware of the needs of distributed teams, but with a majority of their workforce reporting from different geographical locations, there’s a whole new bag of pitfalls to consider. Furthermore, unlike SaaS, enterprises with self-managed environments require upkeep from admins in order to maintain a high level of performance for a high volume of users. Whether the impact is felt at the user level or the admin level, it’s important that we do everything we can to ensure that sub-optimal performance doesn’t become a hindrance to productivity during these unique circumstances.
What’s changed and what’s the same for admins
For the most part, admins are continuing work without too much interruption to their business as usual. When we spoke to our own admins we discovered that since their team was already relatively distributed, most of the processes they had in place were well-prepared to be run entirely remotely.
There are certainly exceptions to this, such as having to upgrade a support site to the latest Enterprise release while your whole team is remote. Some admins will also see an increase in traffic when looking at month-over-month and year-over-year usage patterns. For many organizations, this difference may be minimal, but for enterprise customers whose Atlassian usage spans the entire business, a minimal increase in usage could lead to performance degradation.
When we examined aggregate customer usage patterns from instance reporting, we saw lower usage rates within the first two months of enterprise organizations leaning toward remote work. Usage continued to grow at a gradual pace through the next several months, suggesting that users became more active with Atlassian tools as they settled into their new workflow
It’s difficult to diagnose the exact cause of the increased usage, as it could be due to a number of factors. For example, face-to-face rituals and interactions like backlog grooming, demos, and brainstorming are now being conducted via online methods that may lean heavily on Atlassian tools. The bottom line is that more users are spending more time in Atlassian apps, and this is something all admins should be cognizant of.
Are you seeing usage increase in correlation with remote work? The Data Center team would love to hear your story. Check out our Community post to join the conversation.
There are also features built for controlling the rate of incoming or outgoing traffic, such as rate limiting. With an uptick in traffic for your instance comes potential instability. Maybe it’s CPU spikes from HTTP REST API requests and flooding, or poorly written scripts, or external automation that calls APIs too frequently or don’t cache data.
These scenarios endanger the stability and performance of your instance, and when a single (or a few) bad actors or automated scripts bombard your instance with a massive number of requests and consume the majority of its resources, you’ll experience degraded performance, and possibly even a Denial of Service (DoS). The larger the organization, the more catastrophic this kind of misuse can be.
There are other changes admins may see that won’t directly impact performance. For instance, a team of Jira Administrators who are typically distributed to local offices may shift to a global queue instead of just focusing on the offices they are assigned to. Also, any location-based work will ultimately be moot, so it may be helpful to put these actions on hold and hidden from view.
But we all know a major part of an admin’s role in a self-managed environment is to maintain upkeep to ensure an optimal user experience. Any impact on the user will surely bubble up to an admin, and the more often this happens, the more time is required to mitigate any reported latency.
How you can deliver a uniform user experience
Most enterprises maintain a global footprint, with teams operating anywhere from Tokyo to São Paulo. Even before enterprises embarked on wide-scale efforts to enable employees to work from home, it wasn’t uncommon to see inconsistent performance for globally distributed teams and remote users due to bandwidth issues, and those users’ proximity to the primary server location. Typically, the greater the distance to the primary server, the greater the latency those teams would experience in comparison to their teammates in closer proximity. This latency can be impacted by a number of factors, but one way to combat the issue is by enabling a Content Delivery Network (CDN).
Wait, what’s a CDN?
CDN stands for content delivery network, a globally distributed network of edge servers that cache static resources locally, such as CSS, JavaScript, or fonts. When distributed teams work from a location that’s geographically distant from their server location, it typically means they have to wait longer for a page to be fetched, an issue to open, or a board to load. The purpose of a CDN is to speed up that response time as much as possible by distributing the static assets of these actions spatially, relative to end users.
CDNs are not solely used to combat performance latency for users geographically distant from the server location. Potentially more often, CDNs are deployed by enterprises to improve performance for all users, since higher usage in general can result in peak load on the primary instance. With a CDN enabled, we’ve seen performance improvements of up to 50 percent in Jira requests’ average response times. Evaluating the performance gains a CDN provides will be more important than ever as the number of users accessing the primary instance will likely be higher than ever.
One product-specific example that can also help to reduce the strain on a given instance is smart mirror farms in Bitbucket Data Center. This allows users to cluster mirrors into “farms” grouped behind a load balancer to reduce time spent waiting for build results. Teams can point their builds to a single location (the URL of the load balancer) and add additional mirrors to elastically scale. On top of improved scalability, it also provides high availability; if one mirror in a farm goes down, the remaining mirrors can support the build load.
Interested in learning more about how Bitbucket supports distributed teams and remote users? Check out our blog on the topic.
User management and other security considerations
Protecting remote work is likely at the top of the list for every security team during this time. Just the other day, my wife was on a call with her client when our Google Home chimed in mid-conversation, reminding all involved that the home office still has its differences from the actual office. Suffice it to say, there’s no limit on security risks to consider, and we know that Atlassian tools are no exception.
With the entirety of the workforce logging in from remote locations, ensuring secure access has become paramount. Activities such as authorizing, authenticating, and provisioning user access need to be handled carefully to ensure that the right people have access to sensitive data. It’s this philosophy that’s driven our decision to support leading security standards, such as SAML and OpenID Connect, and it’s even more important in this new environment.
On the other side of the coin, there’s a similar, functional need when it comes to user management. The instant scalability inherent in Data Center products makes it easier to handle more users at the same time, because you can add new nodes to your cluster without downtime or additional licensing fees. User management can impact both security and cost, and it’s times like these that remind us just how critical it is to have clear visibility into which users are accessing what tools, and how often.
What this all means for admins and users moving forward
With distributed work swiftly becoming the gold standard for tech-forward organizations supporting globally dispersed workforces, it’s not farfetched to identify this shift as a significant movement toward a new normal.
As one of our co-founders said in a recent blog post, making a universal transition to remote work with little or no warning is unavoidably disruptive. The best we can all do is focus on adapting so that our tools remain accessible and performant. If we can do that, we can continue to unleash the potential of every team.
To help teams work better together, we’ve also launched our remote work hub, featuring advice on staying productive and collaborative, our collection of free cloud products, and access to integrations from leading Atlassian partners such as Slack, Zoom, G Suite, and Dropbox. If you’re interested in adding more tips and tricks to your admin tool belt, check out the on-demand webinar below.