Taking this as the thread for Jira Ops as well...

> Is there a Jira Ops server option? > Jira Ops is currently available for cloud. We believe the best incident tooling is hosted outside your infrastructure so it’s always available, even if all your internal systems go down.

Once again, Atlassian keeps the good stuff from customers who for whichever reason (usually regulatory) they must maintain systems with no ingress from the internet and therefore cannot use SaaS products.

Atlassian is right, of course, that these kinds of systems which are "above" production should not use the same infrastructure, so that they are completely independent of production in case of wide-ranging production outages. It doesn't matter, regulation is regulation is regulation.

I'm not sure what their market positioning is supposed to be with this product. The value of managed incident response only goes up with the size of the client company. Yet, Jira's cloud offering has a limit of 2,000 users, with early access (read: no production / performance support or guarantees) for 5,000 users. If you work for a large Fortune Whatever company, and even if that company is able to move infrastructure to a public cloud and use SaaS and various goodies, you must operate Jira Data Center to operate with that number of users. Because Jira Data Center isn't Jira Cloud, you don't get access to Jira Ops.

Does Atlassian not think that incident response matters to companies operating under airgap regulations? Or to large enterprises? Does Atlassian think that smaller companies are spending that much time and energy on incident response to warrant this product? :/

> Does Atlassian not think that incident response matters to companies operating under airgap regulations? Or to large enterprises? Does Atlassian think that smaller companies are spending that much time and energy on incident response to warrant this product? :/

I feel your frustration, but that is not the kind of thought process behind such product decisions.

Usually there's some sort of corporate strategy, and if that strategy says "cloud first" or "cloud only", then there isn't all that much that a lowly product manager can do.

And even if it's not just a matter of corporate strategy, these companies ask questions like "what's the market size for a cloud-only product?" and "what is the market size for an on-premise solution?" and "how much more effort is it to offer both?", and if the second isn't large enough to warrant the effort, there won't be any.

From a business perspective, a hosted Saas/cloud product simply is very appealing (revenue stream, homogeneous infrastructure, no offsite debugging necessary etc.), and the business folks always dream of just getting some kind of certification with which they can reach ever more regulated customers.

I've been on the engineering side of sales for a number of years now. The "term" model and deferred revenue is a big push across all public organizations since it bodes well on the balance sheet. The quick death of perpetual licenses for many products that started out in that world is here and rampant. Cloud drives the term model neatly and succinctly. Opex for the SaaS provider is well defined the margins are predictable and scalable. If you like buying your software with a license and having the flexibility my, unfortunate, perspective is your options will continue to be more and more limited unless customers start pushing back.

The other upsell strategy is bundles and ELAs. Sell your customer everything, even if they don't need it. This is done by tossing im the kitchen sink beyond the core product the customer actually wants. It's far better to report sales on products that don't actually sell well on their own and take pennies on the list price vs none at all. I'll refrain from name dropping here but the security vertical is a huge offender of these tactics and I feel as though more production IT/ops products and vendors are now going down this path.

I wouldn't cry for us businesses where a certification or attestation is insufficient and we need on-prem/internal hosting. We usually have the resources to build a homegrown version if necessary.

"Just throw some money at it" becomes a reasonable strategy at a certain scale. Maybe it's time for an open source PagerDuty/OpsGenie offering.

Atlassian wants you to move to their cloud product, it's that simple. Server is a thing that makes their life harder and stops them from extracting maximum revenues from you so you can expect a soft but persistent push to switch to cloud which will get harder over time.

I have no knowledge of what they use to deliver their offerings.

However by keeping everything as SaaS they absolutely avoid all licensing related issues and questions.

Remember that you can keep your changes made to even GPL licensed code secret, as long as you are not shipping code to anyone. So SaaS places no burden on them.

If they were to ship something to a customer then they would have to meet all the requirements of any and all licenses for anything they used.

It seems like most of the point of Jira Ops is to pull together disparate sources of information from different cloud systems (just based on what I see). Just makes me wonder about the server use case.

Jira Ops seems really nice, https://www.youtube.com/watch?v=0uIhtpSMaA4 and has a nice UI

Disclaimer: I work at OpsGenie, and this is the first time I saw Jira Ops in action

OpsGenies Thundra is an exceptional serverless APM tool. It supports auto instrumentation for Java and allows you to debug your exception with your given state & variables at the time (like overops).

Anybody who does serverless should seriously consider using it.

Disclaimer: I am in no way affiliated to any of these companies.

Hi,

I am Serkan Özal, Co-Founder and Lead Engineer of Thundra.

Thanks for your compliment to Thundra :) Are you currently a Thundra user? Have you tried it before? We would like to get your feedbacks. You can contact with me ("serkan@thundra.io") or give your e-mail address so we can contact with you :)

Spolier: There are more advanced features coming soon, stay tuned!

I don't imagine they're making that comment without having tried it before ... which makes this feel like an incredibly scripted response.

including your response and this one?

We use OpsGenie at Sourcegraph. They are great--nice UI, the ability to set rotating on-call schedules, and an API that lets us easily integrate multiple sources of alerts. They were also less expensive by a wide margin than many of their competitors. A great acquisition by Atlassian!

I do wish OpsGenie would use Twilio or someone for SMS messaging so I don't get messages from 20 different phone numbers. Makes it very hard to set a custom alert tone.

Also, iOS doesn't allow you to set a custom per-app alert sound, so I can't set one there that will wake me up when stuff is broken.

Also the UI on their app is marginal at best - multiple taps, no back tracking, etc.

You can't really blame them for the second point, that's an Apple limitation.

I use OpsGenie on Android and I have a specific sound set for the new alert notification. It will also audibly alert me at max volume regardless of my ringer volume or DND setting.

I do agree about the multiple source numbers of SMS alerts but they're hardly the only person who does this.

Can't say I have had many issues with their UI. It's fairly straightforward. The only annoyance for me is ack'ing or closing a number of alerts at once and getting a bit of lag.

Interesting development, I suspected SolarWinds would pull the trigger on OpsGenie first.

Let's see what Atlassian makes of it.

I've found Opsgenie's support and pricing to be good. Hopefully that doesn't change with Atlassian...

You're really comparing two different things here.

Prometheus is great at alerting. So, where do those alerts go? How do they reach the right team, the right person who is on-call? How do they know that Person A is on-call between 8am and 5pm, but Person B's shift starts after that? Oh, wait, Person B is on vacation and asked Person C to cover for them.

Person A, B, and C's manager wants to monitor how the MTTR on all of those alerts is. Where does Prometheus provide that data?

These systems don't do the same things. OpsGenie isn't just an "SMS Gateway".

> Oh, wait, Person B is on vacation and asked Person C to cover for them

You just hit a particular peeve with a sideways glance.

There is a fundamental problem with "person B is on vacation and asked person C to cover". In the OpsGenie world, shift overrides are manual. Only manual. (I consider poking API with pre-calculated data as manual, too.)

What I really want is for OpsGenie to add support for concept of vacation. When someone is going on a vacation, it should be possible to tell the alerting system that during timeframe T, person P is unavailable - now recalculate and reschedule the rest of the roster. But no. They don't do that.

Our guys have asked for this feature during contract negotiations. I have asked it personally from their engineering staff. It's just not a feature they want to support.

you have hit the issue in the head. Even other tools like victorOps suffer from this. Overrides are manual and in some cases, if people switch 4 day on call with 3 day ones. They are getting screwed. There has to be a concept of vacation and automate rescheduling. I am suprised, there isnt a decent tool to address it.

If any Atlassian or OpsGenie folks read here: I wonder if we opensource projects would/could now also benefit from the generous Atlassian licensing for the OpsGenie service? Though "just" an opensource project we do have an "infrastructure" and "tasks" that we would like to be able to "alert on". But haven't do to the pricing.

Hi there. I work at Atlassian. I'll look into this and respond back on this thread. To be completely open, it may be a while before I get an answer since it's just been announced.

For those of you that don't know, Atlassian offers free licenses to open source projects, including cloud and server versions of Jira, Confluence, Bitbucket, and more. See https://www.atlassian.com/software/views/open-source-license... for more info and approval criteria.

+1 to this idea, we would switch to OpsGenie immediately if this was the case (from PagerDuty)