There's a link this week to a post from Brendan Gregg regarding the definition of observability. I don't want to spoil the ending, but Brendan posits that observability metrics alone are not enough.
The post brought to mind a couple of thoughts I have had swimming around my head for a while now. The first thought was how Brendan's post is similar to something I have advocated for a while now regarding the health of your database server and database performance. These are two different things, but people tend to treat them as one. Sure, there are some edge cases where health and performance overlap. But for the majority of the time, they are distinct from each other.
Think of it this way. I'm healthy enough to run a four minute mile, but I physically am not able to run that fast. When users come to you with query performance issues often times you end up looking at health metrics, and not performance metrics. The truth is, you need both to have the complete picture.
The other thought is about those edge cases we all come across. I often see advice given online that a user should buy the most advance thingamabob possible and configure the server like it was a Rubik's cube in order to handle 1% of possible scenarios despite the fact 99% of the time their workload will operate just fine with the current resources allocated.
I can't stress this point enough: Do not architect your enterprise around the possibility an edge case might happen in the future. The best way to prepare for those edge cases, in my opinion, is to use a service that allows for dynamic allocation of resources as needed. This could be virtualization in your Earthed data center, or it could be using a cloud service provider, or perhaps a mix of both.
Know your workload, understand what performance your server can provide, and make sure it is healthy enough to deliver that performance. Don't allocate more resources than necessary for edge cases which may never happen, you are throwing your money away.
Today's Sponsor
Someone is complaining about the database, again. Whatever the bottleneck may be, you know it is not your fault, but it is your responsibility.
Here's a secret for you. The majority of database issues are not edge case scenarios requiring hiring a team of experts to solve.
Common root causes include locking and blocking, poor indexing, or a badly written query. SolarWinds provides a comprehensive suite of cross-platform database performance monitoring and tuning tools for both Earthed and Cloud workloads. And with the recent purchase of Sentry One, makers of SQL Sentry, we can also help with the edge cases, too.
If you haven't tried our products, there's no better time to get started.
Community Links
Evolving beyond password complexity as an identity strategy
Over the past few months we have seen attackers focused on exploiting trust. One possible surface area of attacking trust is getting a hold of your passwords. Troy Hunt talks a bit about the need to move beyond complex passwords as the only means to verify identity.
What is Observability
Besides being a made up word, Brendan talks about how observability metrics alone are not enough information to solve performance issues.
Events
We are moving forward with plans to host Live! 360 this November in Orlando.
Live! 360 brings the IT, Developer, and Data communities together for six days of training, knowledge sharing, and networking. With unlimited access to Live! 360’s five co-located events, you and your team will get the training you need to keep you and your business competitive and future-ready.
Send any questions about the event to me at SQLRockstar@thomaslarock.com
Jobs
Senior Program Manager - SQL Server Big Data Clusters | Microsoft
If you are like me, then the new world of cloud-native computing is an exciting opportunity. And if you also care about open data, scale-out technologies, and data governance then you will want to learn more about this role with the Microsoft SQL Server Big Data Clusters (BDC) product management team.
If this role, or working for Microsoft interests you, get in touch and I will connect you with the right people in Redmond.
Data Janitor Roundup
Keeping track of one cloud provider's data products is a 'full-time job'
More like a few full time jobs, yes. For anyone thinking multi-cloud is a strategy worth exploring, just don't. Pick a vendor and keep as many eggs in that basket.
Create Microsoft SQL Server Instances of Amazon RDS on AWS Outposts
The fact that AWS and Azure are extending their platform into your data center is a clear sign you need to be thinking about improving your skills on one or both platforms. It won't be long before businesses are relying solely on cloud infrastructure masquerading as a local data center.
Announcing Azure SQL Database ledger
This feature has been in development and private preview for some time now, and I was happy to see this announced at Microsoft Build this past week. It's essentially blockchain without the blockchain, and will serve the business needs for companies that need a secure ledger option.
VM Setup — Which Virtual Machine is best for your workload in Azure?
A common root cause for performance issues in the cloud is not having selected the right resources for your workload. And it is a bit confusing, given all the options available and the fact that cloud computing is different than direct attached storage that many are familiar with. This video is a great summary of how to rightsize your workload to an Azure virtual machine.
Perform at-scale, agentless SQL Server discovery and assessment with Azure Migrate
Azure continues to improve in the area of data migrations, making it easier for customers to discover what lurks in their entire data estate and also to migrate workloads to appropriately sized resources. Another sign that you need to be working on your cloud architecture skills.
Sponsor an Issue
Reach thousands of data professionals who care about data, databases, and helping others make their data the best version possible. Get in touch right here.