2024-01-05 - Avoiding single points of failure - Chewing the Fat 📈

[bradystroud]

Hey SSWers,

Lets talk about single points of failure and handovers!

https://www.ssw.com.au/rules/do-you-know-how-to-handover-a-project

[camillars]

Avoiding single points of failure is essential to keep the business moving and running smoothly.

[AlexRothwell]

Single points of failure are important to work on avoiding but the strategy to mitigate the risk must be considered and proportional

For example, throwing out role based protections in response to finding only a couple of people holding the role may speed things up but could also end in disaster

[2fernandez]

I am super confused with this week's CTF. Are we talking about:
"Project handover" or "Single points of failure" ?

Are we suggesting that lack of "Project handover" can cause "Single points of failure" ?

[Dhruv-0987]

The most important thing I learnt was to not mess with Username for one and the importance of secured authentication. Also how returning true instead of false can literally shake the world.

[TobyChurches]

Its definitely important to mitigate single points of failure, especially in large/growing organisations like SSW. Its great to see that task responsibilities are being entrusted to multiple employees in SL, It can be a bit hard to get onto the correct person otherwise 🙂

Still for a number of tasks in SL require a single 'final' approver - Maybe this could be sent to multiple people for approval?

[Calinator444]

I was shocked that the Google presentation was based on a real-life example. Allowing users to login without passwords is the kind of security flaw that drives a company out of business, if not resulting in a hefty lawsuit. Dropbox must have handled the situation very tactfully to stay afloat afterwards.

[JimmyChenSSW]

Ensuring that someone is designated to support you during your absence is crucial, and we strongly advocate for establishing this as a standard or rule.

The rules discussed in today's CTF seem more aligned with transitioning responsibilities when someone permanently leaves a role, emphasizing task handovers.

However, for a seamless delegation of responsibilities during a holiday, our emphasis should be on task organization, identifying alternative contacts, and establishing emergency communication etc.

[wicksipedia]

No mention of the bus test which was surprising

[jaydenalchin]

+1 plus the key person risk

[YalalCakirSSW]

A project shouldn't just rely on a couple of people, and the system should operate seamlessly even when key team members are on holidays. To be able to achieve this, a good handover is crucial. Preventing a single point of failure by ensuring that essential information, responsibilities, and processes are effectively transferred between team members.

[jeoffreyfischer]

Eliminating single points of failure is crucial for maintaining uninterrupted and smooth business operations.

[tkapa]

There's been a couple of times - not even just while I've been at SSW that redundancy or handover processes would have been important.

These kinds of things are hard to spot without persistent analysis, or until it's too late.

[chrisschultzssw]

Important stuff - and good to see SysAdmin related examples :).

[vladislav-kir]

Preventing single points of failure is crucial!

It's great to know SSW's template for project handover :)

[YazhiChen]

Handover is important, and it is always a good practice to perform a proper handover when you are going to be on leave, ensuring that you do not leave the team or other people in a difficult situation.

Even if you are working on certain PBIs and are suddenly assigned to client work, it is always good practice to document what you have done and what remains. This way, the people taking over do not have to chase you for a handover.

[cvionnet]

Nice to have the 8 steps for the project hand-over

[jerryluossw]

I like 8 steps for a good handover very much.
I totlly agree avoiding single points of failure is so important, but I think we can't have redundancy in everything.
Redundancy means higher costs (Much more expensive than single points), we need to think about the balance between redundancy and cost.

[Aibono1225]

It's good to learn from Google's example about how serious a single point of failure can be.

[jimmidier]

IMO dealing with single points of failure in a sociological context can be generalized in 2 categories,

• Readiness-based business - things like this rely on information propagation thus can be accelerated by reducing duplication cost, typically by digitization
• Completion-based business - things like this rely on a distributed system so a consensus mechanism helps in this case

It's like a metaphor of a software system except the latter has the natural advantage of being natively digitized.

[Levijj22]

Great CTF!
Loved the video! Really makes you think about how one mistake/decision can collapse an entire system.
Scientists are saying we could be facing a complete internet outage (for up to a few months) in 2024 due to solar flares, what's our backup then? 👀

[babakamyljanovssw]

Loved the rule for handover the project. Following the steps and template from the rule definitely helps team's work to be smooth. Project handover can mitigate Single Point of Failure, but there is a always chance for SPF, and it is crucial to establish strategies for it.

[tino-liu]

Great Rule! Smooth handovers are crucial to a company's reputation as well.

[sethdaily]

Very important. This should be in everyone's mind when working on processes

[jakebayliss]

My project is especially key person dependent (on the client side). When they are away we have to push tasks back until they return. It's not great. But it seems like this is something SSW is good at as we have always have many devs to a project. Good documentation is key

[bahjatalaadelssw]

Single point failure occurs with poor documentation which is a result of complex processes not allowing for collection of information to maintain a chain of accountability

[Freego1783]

Knowing how to handover a project is really important. I witnessed the acquisition of one company by another and the transfer of a project from one development team to the other. No effort was made, and it was very complicated for the buyers to continue the project.

[stephanfakossw]

It's super important to avoid systems and processes failing because of lack of redundancies .

[piers-sinclair]

Imo, cross approvals are the golden ticket 🎫

[danielmackay]

single points of failure are important to keep in mind when dealing with both systems, people, and processes. This CTF is a good reminder to keep any eye out and address any that we come across.

[Marxoz]

I think the golden rule should be that a process / approval does not sit with only one person. Having 2 tiers is probably also a good idea. One tier that does the approvals / ownership, and another that is available only for emergencies. Having this documented is also important, so that you know who to contact, but also so that the person who is going on leave knows what he needs to handover.

[BrookJeynes]

It's important to minimise the amount of single points of failure on any given projects. Having good handover docs and easy handover processes can help with this.