With the recent capability of setting retention period for Log Analytics data per table a lot of new possibilities of managing and retaining your data pop-up. A common scenario is that you may have a lot of performance data which may be logged every minute or even every 10 seconds. You need that data in such short intervals in your Log Analytics workspace only for the past month or so but you do not need such granularity for older data. At the same time it is good to have some summarization (aggregation) of that data for longer period due to compliance, analysis, etc but there is a cost associated when you retain a lot of data for longer period. By using serverless and the new per table retention capability now you can achieve this and save cost. In this blog I will show you how you can achieve this with simple example.
Lately I have seen some questions and discussions that I have also been involved around which management services/tools should be used when you are doing multi-cloud. Before diving into that area let’s first dive into the multi-cloud thingy. RightScale has report for year 2019 called STATE OF THE CLOUDREPORT that give us what is the current state of companies in that area. If we look at the report we will see that multi-cloud strategy is rising but if we look in the details the strategy of having multiple public or private clouds is actually starting to decline, slightly but still decline. I think that decline will continue over the next years. For me it makes sense if you have the bigger part of your cloud workloads in a single public cloud and may be some small part into another public cloud. My opinion is that it is better to put your bets into a single public cloud. I do not think there are much of benefits if you do multi-public cloud strategy. As for putting workloads on-premises the hybrid cloud strategy I think will be still valid at least for the next 10 years. With that said never the less there are still companies that have multi-cloud strategy with multiple public clouds. And this brings us back to our topic. You have probably heard similar questions like:
As I still see people confused and not informed about OMS I’ve decided to write this blog post and lay my thoughts around the end of OMS.
If you are not sure what is OMS that is ok and you do not necessary need to know but if you are curious you should check out my blog post What is OMS and a Brief History of It. At Ignite 2018 Microsoft probably announced the last major change related to OMS. This was announced on the Azure blog and at Ignite session. To put it in short Log Analytics (which is in 99% what people refer to when they say OMS) is part of Azure Monitor. In Azure Monitor you might see it being called just Logs for short. But besides that there are actually way more changes some of which happened at Ignite others were happening for quite some time.
Recently I’ve came into the following situation. I needed to store a json as Azure Automation string variable. To do that is easy as you just need to pass the json as text by first escaping it. That is easy peasy when using PowerShell. But what if you want to pass the json as object parameter via ARM template parameters file and do the escape completely within the ARM template. Apparently that is possible as well and I will show you how.