Statistics in the BIT Portal
Statistics in the BIT Portal
Statistics. Because if something isn’t monitored, it doesn’t exist, right?
From an earlier BIT blog:
Timeseries data is, in principle, telemetry generated at regular (or irregular) intervals. It reflects the state of something at a given moment. Examples include memory or CPU usage, the number of web requests processed, temperature measurements, or even when doors open.
This data can produce interesting graphs, as seen, for example, by our customers in the BIT Portal. We can observe things like a disk filling up, the busiest times of day or week for traffic, or even debug an issue by looking for notable patterns in the data—such as realising that it’s not a web server that’s overloaded, but a database.
In short, useful information
So useful, in fact, that we’ve been providing customers with a wide range of statistics and graphs in the BIT Portal for quite some time now. These statistics are automatically generated as soon as you use our service(s), and we’d like to give them a bit more attention.
Which statistics are available?
The following statistics are standard in the BIT Portal when you take up the relevant services.
Data Centre
When purchasing rack space, statistics on power consumption and power pricing for the relevant racks are available, along with data centre temperature readings.
Hosting
For each virtual machine, a statistics page is available showing CPU usage, disk I/O, network traffic, and in some cases, memory usage.
This is a limited dataset visible from the hypervisors without requiring access to the VM itself, so it’s available to all customers. Memory usage display does not work for Windows machines.
For those with a Virtual Data Centre (VDC), alongside individual virtual machines, we also provide an overview of CPU, memory, and storage quotas, as well as network usage per calendar month, broken down by VM, VLAN, and network classification.
Jitsi
Customers of the Jitsi service have a dashboard that shows how many people are active in how many different video calls at any given time.
Load Balancing
Customers with one or more load-balanced Virtual IPs (VIPs) receive a dedicated page per VIP displaying requests, connections, bandwidth usage, and the availability of the Virtual IP.
Each Virtual IP consists of one or more connection pools with one or more server nodes—think separate pools for IPv4 and IPv6 traffic. Per pool, there’s a page with graphs showing connection and request volumes, both in total and per node, as well as the availability (or lack thereof) of the pool and its nodes.
Network
All our customers receive two standard pages displaying global latency and packet loss statistics for the BIT network. Even if you’re not a customer, you can view this data at bit.org.
Additionally, there are pages showing incoming and outgoing traffic for every port on each type of connection you’ve purchased, whether it’s our core network or an NDIX connection, plus a page aggregating inbound and outbound traffic across all ports.
Storage
Customers of shared storage (CephFS) have a page per share showing its usage, both in absolute numbers and as a percentage of the quota.
Security
Now, that’s a lot of data, you might be thinking. Is it all secure? Can other customers access my data?
This data (and much more), is collected centrally within BIT in time-series databases that are inaccessible to customers or the outside world.
Because these central databases contain not only customer data but also our own internal data, the specific customer data displayed in the BIT Portal is filtered.
Only the data relevant to the portal is stored on a separate server, with each customer’s data split into separate databases. This ensures that each customer can only access their own data, even if someone were to gain direct access to the database rather than just the dashboards.
This way, we keep customer data (and, of course, our own data) properly segregated.
By: The people at BIT
