Service Desk Toolbox
A multi-tenant troubleshooting console. An engineer picks a customer, picks an application and looks a person up - instead of logging into a different vendor portal for each one. Alongside it runs a unified network dashboard that folds two different hardware vendors into a single fleet.
Built for Service desk teams supporting many customers across several platforms
This one runs on live business data, so there is no public demo. A walkthrough on a call is the next best thing, and we are happy to do it.
60 s
Fleet poll interval
5 yrs
Device history retained
Deployed configuration: how often the fleet is sampled, and how far back the recorded device and outage history goes.
Answering "why can this person not print" meant opening several vendor portals, remembering which customer used which system, and knowing each product’s own quirks. The network side was worse: two different vendor clouds, neither of which keeps history long enough to answer what happened last Tuesday, and no single view of which sites were actually up.
A shared shell with one provider module per application, so a new system drops in without touching anything else. Behind it, a standalone poller records the history the vendor clouds throw away - device state, WAN quality, outages - and the dashboard presents both vendors as one fleet, grouped by internet provider rather than by who made the hardware.
Pick a customer, pick an application, search by name or email. Membership, plan, credits, contracts, passes and team details come back in one view - and opening a second application for the same person carries the search across instead of asking again.
Sites from two different network vendors appear together in one dashboard, each tagged with where it comes from and filterable to one or the other. Internet provider health folds both vendors’ sites into a single group per provider, with combined uptime and outage figures.
When the poller was down, or was up but had lost sight of the fleet, that period is recorded as a monitoring gap and excluded from uptime rather than blamed on the equipment. If one poll flips most of the fleet at once, that is treated as a gap too, not as a fleet-wide outage.
A standalone process records device online and offline state, WAN quality samples and durable outages continuously, over windows from an hour to a month. Every device shows how long it has held its current state and when it last dropped.
Each integration’s live status, response time and last-24-hours uptime on one page, recorded continuously by a background monitor rather than checked when somebody complains.
Each application is a provider module behind a common interface - authentication, token handling, lookups and error shape. Adding the next one is a new module, not a change to the shell every other system depends on.
Chosen for how long it will stay supportable, not for what was interesting that year. Everything below is in production in this project today.
The work this involved
Service Desk Toolbox started as somebody describing a process that had stopped working. If that sounds familiar, that is all we need to start.
Discovery and written scope - free, and yours to keep