Oracle Management Cloud is a new Oracle Cloud offering launched during Oracle Open World 2015, of which I am very very interested in. Currently you only can get a paid subscription (no free trail available yet) so the following here is to give you an overview of what it is, based on different sources from the internet and some Oracle Sales material. Strangely enough, I think, if I see for example the view count on the available YouTube movies, it hasn’t got that much traction yet.
So why am I so enthusiastic about this particular Cloud offering?
A bit of personal history
When I started my IT carrier as an Oracle database management person (20+ years ago), and later on wondered also into other specific areas like managing operating systems, hardware, application servers, programming language environments, (cloud/virtualized) architectures, etc. (while focusing on management, performance tuning and troubleshooting), I realized some of the following:
- Its good to have a single birds eye view (it starts with a good dashboard)
- Performance is about the whole stack (most of the time its not the database that is the culprit)
- Good instrumentation is key
- Having all information at hand, that is the ability to see everything in context
- Having the knowledge to see it all in perspective (connecting the dots)
From bag of tricks toward centralized OEM setup
To clarify some of the pointers. When I started at my first IT job, I had to manage distributed Oracle 7 environments on Windows NT 3.0 (I am not kidding you) between Rotterdam (The Netherlands) and the Ipswich (UK). In the client-server topology of that time, it means that I had a bag of tricks to, via Windows SDK and other tooling, to remotely manage (e.g. start or stop) database instances, Windows services, etc., in the UK; hundreds of DBA SQL scripts, and other solutions that helped me do stuff (all from the command line), Don’t forget, at that time (20 something years ago), Windows NT 3.0 wasn’t much more than a glorified Windows 3.11 machine.
Later on, working for a different employer, I remotely managed (on my own) several business critical, some simple, database instances. So I almost immediately embraced Oracle Enterprise Manager 10g when it came out (10-15 years ago?), and traded my bag of tricks, small own build web based SQL framework (HTMLDB/XML SDK Servlet) for this new OEM web based solution. I had now a centralized view of all my environments that I had to manage and more.
Somewhere in my very old email folders, I have a picture that I created for my manager at that time, that displayed a remotely managed topology / environments overview, which could be easy to setup with a CD at hand that contained OEM agent software. Installing the agent software on site, via this software CD, enabled connecting the database and other Oracle environments, up with our central “OEM hub”. Its funny when I think about it, because I still can remember a emotional debate with an EMEA Oracle person at that time, trying to convince him that OEM could be used for more then local database management on premise (how we call it today).
Interpretation, knowledge level and context
Even if you are knowledgeable in (remote) Oracle database management and have a tool like OEM at hand, it doesn’t give you the ability yet to “connect the dots” if something goes wrong, like performance problems. You must be willing to acquire knowledge about the stuff on the other side of the fence as well. If you do, you will be more valuable and effective for your business. You will have the ability to function as a bridge between multiple disciplines (management/developers/database management/hardware, network guys). In all its an enriching experience and, for example, it makes you a better database guy (or DevOps+ guy to follow the buzz a bit).
A long time ago, this topic was a bit touched via the “DBA 2.0” initiative, to promote Oracle Enterprise Manager (OEM 11g?). My problem with this was that it was focused on the tooling and that it might become a dangerous situation if someone without the proper training and knowledge (and knowledge of the business) just would follow blindly the advise given in OEM and start to “push on buttons”. A tool like OEM has no insight in how your business supporting database has been set-up and why. To give you an example, just creating “missing indexes”, according to OEM, or setting a database cursor parameter, might slow down processes, you didn’t know about yet (but you will, when people are starting calling you…). Also, I solved a lot “database related performance” issues, by just fixing an issue outside the database like removing entries of old commissioned namespace servers…
Unified single point of view complexity
A while ago, during a job interview, a younger person (seem from my perspective – and who is not nowadays 😉 ), asked me if I could tell him where all the specific Oracle RAC log files where. My first response was, but I didn’t (job interview and all), if he knew where all the WebLogic / Oracle Application Server logging where situated. That said, its almost impossible nowadays to know where all the key log, trace, alert files are which one should examine to be able to see the context of it all, that is, all factors that contribute nowadays in a modern environment (lets keep it simple…): client machine, client software, middle tier (hardware/software/versions/patch info/operating system/oracle network/connection pooling/clusters), network connections, in between infrastructure like switches, routers, DNS/name servers, security implementations, load balancing hardware, failover architectures, database (software/operating system/patches/hardware), SAN/NAS hardware, SAN/NAS vendor specific I/O solutions, connection setup to the SAN/NAS environments/etc., etc., etc.
And we are not even talking yet about end user behavior (“Do you know that the software was not build for that purpose?”, “That’s the wrong button b.t.w.?”, “You shouldn’t quit the browser based application by using the [x] – this will create a lock in the database”), IT knowledge level of everyone involved on all the different levels (“You have forgotten to setup this bit with parameter x”, “The patching off the database caused that middle tier SQL*Net connection to fail – incompatible security connect methods”, “Altering the OEM server node name, caused failures across the board regarding SSL connections”), etc.
Lets say you encounter someone on the street, a person you know, and she or he is sweating a lot. Do you think he might be sick or do you ask him how much miles he has run just now? Think about it. Both might be true, although probably not at the same time (link: some causes that make you sweat). The ability to assess the situation correctly is all about context and the knowledge you have. Interpreting all the info available correctly is hard, really hard, and still a mayor problem/risk for your business driven environments to correlate cause and effect correctly within the context it occurred.
This is where the “Oracle Management Cloud” offering comes into play and could help solve some of the issues mentioned…
Oracle Management Cloud
A small disclaimer here: this is what I think it currently is, without having the (trial) access to check up on the service offering “as is” and the information out there on the internet, which is not that much yet (at current)…
In principle the Oracle Management Cloud is offering three cloud services:
Based on the “Oracle Management Cloud brief” document (summarized here):
- Oracle Log Analytics Cloud Service stores and correlates machine data from many different applications and devices, from smart phones to mainframes.
- Oracle IT Analytics Cloud Service allows you to step back and assess macro issues. It enables you to also evaluate systemic problems impacting application performance and use event analytics to fix faulty or misconfigured components.
- Oracle Application Performance Monitoring Service provides end-user and application performance monitoring across a wide array of technologies with minimal up-front setup.
Via a software agent package (and enhanced OEM agent or even separate agent product?), you connect your environments to the “Oracle Management Cloud” environment.
This will enable the Management Cloud to get hold of all the trace, log, alert, etc. files it can consume out-of-the-box formats. You have the ability to consume even more formats via creating your own “parsers” based on regular expressions.
At current Oracle supports 50 or so different formats (and will create more in the future).
All these files, all this info, end up in one unified big data environment.
So now you have everything you need in one unified environment, no need anymore to search for all those different log, trace, etc files to see them in context, that is, it is now all under your finger tips to check out in the log analytics service offering. Also the files / content can be grabbed from all over the place, your local site, your Amazon, Oracle Cloud or Azure environment. And because “Oracle knows Oracle” as I always mention, it (should) know(s) where all the important info is, where to gather it from, which directory…
As mentioned in The Log Analytics documentation:
Using Log Analytics, you can:
- Explore logs specific to the application experiencing a problem
- Analyze and explore log data efficiently
- Gain business and IT operational insight from log data
- Rapidly derive value from logs
- Troubleshoot problems from end user performance issues down to logs
The video mentioned in the documentation gives you a feel of what it could do for you.
It doesn’t give you the overall solution(s) yet when dealing with (performance) problems yet if you don’t have the knowledge to interpret what is in front of your eyes, nor the correlation ability. You still need the knowledge to correlate, for example, which database is having what kind of impact on which application server (and/or vise versa), but it will provide you with a lot of information in one place without having to actually connect to all the environments, figure out where the (trace/log/etc) data is.
You will have the ability, among others, to:
- visualize data
- search data
- see trends over time
- filter data
As mentioned in the IT Analytics documentation, here you can (for database and middle ware only?):
With IT Analytics, you can:
- Analyze data by various dimensions
- Correlate data across charts
- View the top X by a specific measure
- Forecast performance and resources, based on historical and real-time trends
- Save and share your analyses
- Build dashboards to present custom views of data, including IT asset inventory, data center performance, service level agreements, and so on
So in this service offering, Oracle takes care of some of the heavy lifting regarding giving analytics insight into:
- Resource analytics, like:
- IT resource utilization resource consumption,
- Compare data,
- Giving insight about future resource planning
- Performance analytics, like:
- Out-of-box analytics for most common resource bottlenecks
- Visibility into configuration profiles of tier-specific assets based on performance
- View the impact of shared resources
- Insights due to being able to create correlations
- Create Dashboards (Yeah!!!)
- Allows you to build and publish custom dashboards to key stakeholders, e.g. line of business executives or heads of infrastructure or operations
- Data Explorer capability
- Browse or perform advanced analytics on the vast array of operational data and over the long term
The following video mon YouTube gives you a feel of what it could mean to you
Application Performance Monitoring (APM)
As mentioned in the Application Permance Monitoring documentation, here you can:
With Application Performance Monitoring, you can:
- Rapidly isolate application performance issues
- Drill down to related logs in context of a problem and find its root cause
- Gain end-to-end visibility into the performance of your application across all tiers
- Monitor end-user experience
The video mentioned in the documentation gives you a feel of what it could do for you.
Troubleshooting with Management Cloud
Have a look at the following YouTube movie and decide for yourself if it would be “a nice to have” service.
Observations (from afar)
Be aware that all those (Oracle OpenWorld 2015 based?) YouTube movies start with the general Oracle disclaimer, so without a (test) trail account I can’t verify what is, what will be there currently in the Oracle Management Cloud.
I really liked the dashboard as seen [here], but have no clue, because it shows application, host, storage, database and network components, with which service / tool, this dashboard can be created.
Maybe its part of something else or a main “Management Cloud create dashboard” section.
The current available documentation is a bit sparse when it comes, for example, creating dashboards or other functionality.
That said, I would be interested also in some info about, for example:
- Can I create or access REST or other API functionality to embed into a DevOps architecture to signal, for example, “deployment didn’t succeed (disk full)”?
- How long can one keep the data, due to SOX regulations and/or can I can get the correlated info out to my system (security analysis purposes)?
Anyhow. I like the offering and of course it is only launched a short while ago, so take that into account as well. Just like a tool like Oracle Enterprise Manager, its is as good as the guy who is using it, but will pinpoint you probably more easily at the issues at hand, give insight in old and predict a bit of what might be happening in the future.
Have a look at it and I really hope it will get more traction (and a test trail possibility) then currently is shown per view rate on those YouTube movies (a few hundred at best).
Some references on the internet that I found, that gives you a starting point regarding the Oracle Management Cloud:
- Oracle Management Cloud – Home page
- Oracle Management Cloud – Main Cloud page
- Oracle Log Analytics Setup – Overview
- Oracle Management Cloud – Brief Introduction
- Oracle Management Cloud – eBook
If you look up “Oracle Management Cloud” be careful not to end up on “Oracle Managed Cloud Services” offerings or “Order Management Cloud Services” page information, which are completely different service/product offerings compared to “Oracle Management Cloud”.
Whats in a name, right…?