Hosting
Hetzner
TaskFlow instance hosted at Hetzner's main data centre in Johannesburg have a standby instance at Hetzner's data centre in Cape Town. Hetzner's data centres are all equipped with CCTV cameras and access is granted to Hetzner employees via biometric access. We try our absolute best to secure your data and prevent unauthorised access while keeping your data accessible whenever it is needed.
| Info | ||
|---|---|---|
| ||
From 1 September 2019 all TaskFlow databases are hosted on Azure Cloud. |
Azure Cloud
If you are a SaaS client, your TaskFlow instance is hosted with Azure in the South Africa North (JHB) region with backups being made to South Africa South (CPT) region. Azure has more regions than any other cloud provider and they provide complaint and resilient solutions.
On Premises
We do By default, all clients are hosted inside one or more of the Terraco data centres located in Johannesburg, Cape Town and Durban. These servers are managed by TaskFlow and we ensure high uptime as well as data replication between these data centres to ensure the best experience for our customers. We use a hardened operating system and are up to date with all of the latest security best practices.
We partner with CloudProx, Hetzner and Microsoft Azure to ensure full data redundancy and always store backups in different locations should a data centre become completely unavailable. All data remains within the borders of South Africa.
We accommodate clients who want to host TaskFlow on premises , in which case it is the client's responsibility if there is a good reason for this. In these cases the client needs to provide us with a server . On premises clients have the responsibility to restrict access to the server and are also responsible for making backups and storing the backups securely.and is responsible for securing the server. It is also the client's responsibility to perform hardware maintenance and make daily backups.
Availability SLA
Our server infrastructure will improve over time. At the time of writing we are able to provide a 99.9% uptime to our clients. This excludes the on premise clients since we cannot control all the variables, such as power supply to the server and network uptime.
| Availability % | Downtime per year |
|---|---|
| 99 | 3.65 days |
| 99.9 | 8.76 hours |
| 99.99 | 52.56 minutes |
| 99.999 | 5.26 minutes |
| 99.9999 | 31.5 seconds |
Backups / Disaster Recovery
We keep 7 days full backups for each TaskFlow database. Backups are replicated ( via a secure channel ) between regions as well as to a physical location outside of Azure Cloud. Backups are made at 02:00 every day. Backups are not sent across the boarder - i.e. the data remains in kept within the borders of South Africa.
Hardware failure
- All of our servers' hard drives run in a RAID 1 configuration. In case a server fails completely due to hardware issues, we will be able to spin up all of the clients on the counterpart slave server within 4 hours.Hard Drives
- RAID configuration ensures data loss is impossible in case of a single drive failure
- Entire drive snapshots are made and securely transferred to a different region
- Memory
- Multiple memory modules ensuring no disaster for single failure
- Power Supply
- Dual hot-swappable supply ensures uptime even if one power supply fails
- Network
- Two physical network cards ensures uptime even if one card fails
- Interfaces are set up in a Linux Bond ensuring automatic fail-over
- Multiple breakout points to the internet are provided via BGP
Disaster recovery
- RPO (Recovery Point Objective): Clients will never lose more than 24 hours worth of data. This will only happen when we are unable to restore a slave spin up the High Availability server and have to restore your TaskFlow instance from the previous days' backup.
- RTO (Recover Time Objective): Clients will be offline for a maximum of 4 hours in case an entire data centre becomes unavailable. This is due to configuration needed in order for the High Availability server to be fully up and running
Servers are constantly monitored and TaskFlow system administrators are notified whenever crucial events occur.
...
Login details are hashed using industry standard hashing techniques. Not even TaskFlow staff has access to your passwords. In the event that a client loses their password, the only option is to send password reset instructions. When logging into TaskFlow, if you're a SaaS client, your data is always transmitted via HTTPS. You can verify this from within the browser by making sure that a "lock" icon is present.
...
All code changes made to the TaskFlow framework is evaluated and reviewed by TaskFlow engineers before it is deployed to our clientsthrough a peer review process. In addition, our automated testing platform performs unit tests and code auditing and provides our engineers with reports that are used to determine if changes introduce any bugs or regressions.
Changes are carefully written up and compiled in a new version release document which is distributed to clients prior to deploying new versions. Opportunity for feedback is given and our engineers ensure that upgrades always result in a better product for the user.