Skip to content

Commit

Permalink
Merge pull request #98 from exasol/nc-update-monitoring
Browse files Browse the repository at this point in the history
Nc update monitoring
  • Loading branch information
SWEXA authored Nov 13, 2024
2 parents 4f85ae2 + c53eb8c commit 9331931
Showing 1 changed file with 29 additions and 38 deletions.
67 changes: 29 additions & 38 deletions Support-and-Services/monitoring-service-faq.md
Original file line number Diff line number Diff line change
Expand Up @@ -2,35 +2,49 @@

## Background

Exasol's current Monitoring Service is built on a Nagios setup based on the legacy XML-RPC API. The new Monitoring Service is a generic implementation no longer dependent on an external API. Instead, locally installed monitoring agents on each Exasol node collects relevant monitoring information.
Exasol's Monitoring Service is a platform hosted and managed by Exasol which allows us to receive data about the performance and stability of your databases. This data is then used to provide better and faster support. For customers with Platinum Support, connection to the Monitoring Service is required for 24/7 alerting and incident management.

The advantages of the new monitoring service are:
Connection to the Monitoring Service gives you the following benefits:
1. Faster problem identification as support is able to view a subset of logs before they are sent to Exasol
2. Better performance analysis by Exasol experts using pre-defined Grafana dashboards based on best practices
3. A quarterly Usage Report as a PDF with an insight into your database usage and any potential findings that require action.
4. (Platinum Support) 24/7 Monitoring, Alerting, and Incident Management

## How does it work

The Monitoring Service is a generic implementation using open source tools. On each node, Exasol Support installs monitoring agents which then collect relevant monitoring information from the system and the database. The Agent then sends the information to Exasol's central harvester and transforms and saves the data. Exasol Support has access to a variety of Grafana dashboards which are built on top of this information for speedy troubleshooting and incident alerting. The monitoring stack includes the following features:

- no single point of failure
- easy and secure connectivity
- fast deployment
- scalable architecture
- modern stack
- web GUI (planned for future use)
- ready to be used with the new v8 release (ETA 2023 - Q1/Q2)

## Explanation
## How do I get connected to the Monitoring Service
If you are interested in getting connected to the Monitoring Service, just [open a case](https://exasol.my.site.com/s/create-new-case?language=en_US) with us expressing your interest. During the processing of the case, you will need to work with your internal teams to allow your Exasol clusters to connect to Exasol's host harvester.exasol.com via the following public data gateway ports (TCP):
- 9092
- 10016
- 10019

All nodes must be able to resolve the hostname **harvester.exasol.com**.
<br /><br />
After the internet access is available, Exasol Support will set up a meeting with you to install and configure the agents if there is no VPN already configured. Installing the agents does NOT require a downtime.

### What are the monitoring agents?
## What are the monitoring agents?

The monitoring agents are created using [Open Source Telegraf Server Agent](https://github.com/influxdata/telegraf). Updates are provided on a regular basis and can be applied by our support staff.

### How will the new Monitoring Service be installed on your Exasol environment?
## How will the Monitoring Service be installed on my Exasol environment?

Exasol will install the monitoring agents on your clusters. These agents require root access and run inside the Exasol cluster process namespace.

The Exasol cluster where the monitoring agents are installed must have access to Exasol's Data Gateway via the Internet. All data sent by the agents is encrypted.

### What data is collected?
## What data is collected?

The following information is collected:

**Note**: Just like our current monitoring solution, Exasol does not collect any personal data.
**Note: Exasol does not collect any personal data or have access to user schemas.**

- Exasol database statistics 
- Schema "EXA_STATISTICS"
Expand All @@ -55,7 +69,7 @@ The following information is collected:
- Rsyslog (log stream)
- Exasol logs 

### How is data transferred?
## How is data transferred?

Once data is collected by the nodes inside the cluster, it is converted into the Influx line protocol and shipped to our Data Gateway (harvester.exasol.com). The connection uses SASL over SSL. Data will be sent to the monitoring stack via the Internet.

Expand Down Expand Up @@ -83,43 +97,20 @@ Monitoring agent certificates used by the Exasol monitoring agents (can be downl

On top of that each Exasol cluster is using a unique user + password combination in order to authenticate at harvester.exasol.com only if certificates + user + password do match monitoring data will flow into our monitoring platform.

#### SOCKS5 Proxy support
### SOCKS5 Proxy support

If direct internet access for the monitoring agents is not allowed, data can be transferred through a [SOCKS5 proxy](https://en.wikipedia.org/wiki/SOCKS#SOCKS5). The SOCKS5 proxy must be able to resolve harvester.exasol.com and to access the TCP ports listed below. The monitoring agents installed on the Exasol hosts will then send their sensor data to one single port on the SOCKS5 proxy.

The agent only supports SOCKS5, no other SOCKS protocols are supported.

### How is data stored?

Data is stored at Exasol on an on-premise system. As with our current monitoring solution, the data itself is unencrypted but the underlying hard disks are encrypted.

### What’s next?

1. Existing customers need to allow their Exasol clusters to connect to Exasol's host harvester.exasol.com via four public data gateway ports (TCP):
- 9092
- 10016
- 10019

All cluster instances must be able to resolve the hostname harvester.exasol.com.
2. Exasol will install the Monitoring Agents on the existing Exasol clusters.

A downtime of the database is not required.

The current and new monitoring solutions will run in parallel starting from 09.01.2023 - 31.03.2023.

### When can it be installed?

Release date: 09.01.2023
Retire date of existing monitoring service: 31.03.2023
## How is data stored?

#### Rollout
Data is stored at Exasol on an on-premise system. The data itself is unencrypted but the underlying hard disks are encrypted.

Exasol Support will get in touch by the release date with you and plan the rollout.
## Will I get access to the Grafana dashboards?
Currently, access to the data and the dashboards is limited to Exasol Support. We are evaluating if and how to make this available directly to customers. Each customer will receive a quarterly Usage Report with screenshots from the dashboards and key findings on performance and usage trends.

#### Requirements

- Firewall settings must be adjusted so that the host harvester.exasol.com can be reached via the aforementioned ports.
- A session is scheduled to install the new monitoring agents if no VPN between Exasol and your environment is existing.

## Additional References

Expand Down

0 comments on commit 9331931

Please sign in to comment.