We've been going a bit off-topic in the latest months, but in September we're finally back on track. This month's theme should especially catch the dashboarding enthusiasts' attention.
Following the inspiration from our unstoppable Dynatracer @zietho, we'd love you to show the fanciest explorer tile-based dashboards, your coolest and most insightful metric expressions, or simply beautiful visualizations you're proud of! Share it any way you prefer, so other members could easily check it out.
As always, we've prepared exciting rewards for everyone who'll submit their unique answers down below. You'll get a unique "Data Explorer" badge, 100 bonus points, and who knows... Maybe you'll learn something exciting from other participants 😉
That's a good challenge, will watch it to steal ideas, 🤣.
Mine is not fancy, nor complex queries... the beauty here is that my client was looking for a way to monitor their http requests result done thru a Synthetic Monitor. They know that we can do calculations in post execution scripts in the Synthetics, but we can not store values to compare in the following executions.
So the idea I had was to add a new event to their Synthetics, this time, calling the Dynatrace Metric ingestion API, so we can have the results history and watch them in a Dashboard and even do metric events monitoring over it.
I am using the Synthetic entity as dimension to the custom metric, so I can attach the alerts generated to the client MZ. The SM variable there in the script is collected parsing a json, in the previous event, the one client is using to test their app.
So we have two monitoring in one. The app availability tested by Synthetic, and the data value monitoring, with history, done by the custom metric injection.
I guess it does fit in this challenge.
Thank you @dannemca!
And about "stealing ideas" - one of my friends who conducts training, always says at the beginning to the attendees:
"If you will find something really valuable today - you can "steal it" and pass it on. As long as you share your knowledge with good intentions - it's not a crime. 😉 And if you add my name to it, I'll be glad. 😀"
Maybe it's not the most breathtaking visualization, but an attempt to discover a new use case for Dynatrace dashboards. Here in the Dynatrace Doc team, we're experimenting with the Dashboards and Extensions 2.0 framework to add the release notes to the product in a way that it's detached from the cluster code. A very basic extension with a dashboard JSON as an asset does the job. This approach would allow us to ship the release notes when the version is released. We're thinking about the automation that would build the extension based on the same source from which the online release notes are cgenerated. And here's a mock. One of the great things about it is the ability to link directly to places in the Dynatrace UI.
@Kenny_Gillette what do you mean by "And add anything that aren't in release notes?" Our goal is to transparently communicate all the changes. If in the end, the release notes make it to the product, the content scope should be the same, but with the advantage of following the direct links, rather than reading "Go to Settings > Feature > Settings" menu cascades. Let me know what you think 🙂
There are items that are released that don't make release notes. Or they don't make in release notes right at first. Perfect example - data explorer (templates) was not in the release notes on day 1. It was added later. I have also had discussions with people at Dynatrace that even told me not everything makes in release notes. Hope this helps.
Thank you @jaroslaw_orlows!
Any automation that simplifies the work is desirable, but having this one would be great. Fingers crossed for this experiment. 🤞
Challenge accepted! :). I am working with a customer who's utilizing the dashboarding capabilities nicely. A mix between custom and predefined tiles to customize their landing page. A holistic view of the environment while addressing the most common questions.
I'll point to "cool" metrics. In this case it's about a synthetic monitoring made on real mobile phones, in this case an iPhone 6S. It's a 4-step transaction, with timing data and additional metrics grabbed from the phone:
Not something new but for my customer I have been using the licensing dashboard that is capturing the data using the license extension. The dashboard comes from a template that already existed. This dashboard groups up the DDUs, Synthetic, RUM, DEM etc all in one chart
The "Kubernetes Memory & CPU Usage by Workload" dashboard helped us in our k8s cluster to find the right limit/requests for cpu and memory for our workloads. Its interactive and let you filter for your k8s deployment. By default it shows the top 10 wasteful deployments to use our resources wisely.
I want to talk in this one regarding the transformations that become available on Data Explorer, and, specially, for my favorite one that is the parents Transformation.
It solved several use cases where the Key Request needs to be in context in the visualization with the entity they belong to (among other use cases!). This Transformation provides just that, so is really a handy one!
I´d been working on a dashboard to show the impact an outage might have on real users.
This dashboard calculates an average of the traffic for the past four weeks, and, if there is an outage, it makes the math to estimate how many users might be impacted and how many might not convert because of it. We can even calculate a loss of revenue depending on if we have the avg cart value.
I also added visualizations like Errors, Visually Complete, Error page, and real traffic to identify if the outage might be related to any of these causes or if we can see a drop in real traffic.
This is a sensitive dashboard; it is essential to have a synthetic monitor that represents the conversion journey very well and has to be very well maintained. And is very important to communicate to the client that this can only estimate an impact on real users.
I think it is an excellent way of showing clients who rely on our synthetics to measure their applications' availability and the impact that the outages might have.
It is not the same to have an outage during the night, when there is no traffic than having it at the time of the day when there is more traffic.
Availability as a metric is calculated by time, and this can complement to understand the impact of an outage.
This is a journey dashboard created for a customer to showcase the health of multiple components in an application & overview of transaction from Source -> Middleware -> Actual App (and its components) -> Backend.
Couldn't capture the whole dashboard in one snap, hence above snap is a part of the whole :):)
This is a hard challenge as I have so many sweet dashboards.
First, our Dynatrace Admin page. This page elaborates all the details an Admin would need. Granted, its more of a Markdown Page, but still its been a huge benefit for our Admin Team:
Lastly, our customer journey page which provides valuable information as to what Issues/bugs/errors that our paying customers run into. This provides the web development team with insight into their design and how it functions with the average customer. For example, we once reviewed a session replay where a user spent 30 mins paying his bill, he reapplied his card payment details, billing address and everything, the issue was they forgot to supply their phone number which was a required field, but there was no popup directing them to that page. Ultimately, the user finally scrolled up high enough to see that the phone number was missing, which posed the question.... is a phone number really required for payment? When we presented that to the Web Developers they informed us that it actually wasn't, and that they have planned and set in motion a change that will remove that as a required field in the next release.
Huge fan of what you did if it comes to the structure -- the admin has one glance and all the information is here! And I second Karolina's comment on icons!
I have just posted a new Kubernetes cluster overview dashboard in the Dynatrace Tips, I guess it is not so bad:
But I am proud for my next dashboard, where I combined the performance data with busniess data within one dashboard. I used the custom db query results (login, paymant transaction count, payment trasaction amount), SLO metrics and of cource service response time metrics. After an incident (eg, service outage) estimation can be done about the losses based on the trend data. One of my favourite feature is also applied on this dashboard: TimeShift. I love it.
Login data (by authentication channels via db query), Autorization service SLO and login trend (timeshift (-1w):
Response time degradation with timeshift also:
Business data with db query (transaction count and amount) with timeshift:
I have multiple Dashboards,
But in My case my favorite which is common to all customers is the Admin Dashboard.
We can see interesting links, links to other Dashboards...
We are still adding more things on it.
I'm working with a customer who wants to know every release of their mobile application and how fast the customers are adopting the new versions.
Also, they want to know if the new versions are better or not in terms of Requested Errores, Apdex, Duration, etc.
They are viewing every minute how their principal microservices are experiencing degradations or not, and take action before the end users where impacted.
It may not be the most sexy dashboard but a day-to-day faithful companion to manage lambda performance for my client's devops team :
Here is my contribution!
Use case : Build a monitoring for Helpdesk Team to check MES (Manufacturing Execution System) configuration compliance. One MES host per factory. I had to build some extensions to track different datasources and custom needs.