Why Do Companies Tolerate Critical IT Infrastructure Problems for Years?
Asked by
MellowPine47
I'm a general IT specialist supporting more than 2,500 users across multiple countries, and I'm struggling with infrastructure problems that I don't have the authority to fix. Several remote sites have experienced the same serious issues for more than five years: 802.1X disconnects users after an hour or two and often refuses to re-authenticate them, Wi-Fi regularly reports that no network is available, and most locations depend on a single ISP. When that connection fails, ERP systems, file shares, and normal business operations stop.
The company appears to have enough revenue to fund better hardware, SD-WAN, redundant internet connections, or an outside consultant. However, requests for improvements continue to be ignored. When a site goes offline, the team responsible for networking and infrastructure is often unavailable or responds passively while the location remains unusable.
How do organizations function this way for so long? How should I handle the frustration of watching preventable issues continue year after year when the solutions seem obvious but are outside my control?
Answers
Document every incident in neutral, measurable terms: location, start and end time, affected users or departments, systems unavailable, business work lost, probable cause, and recommended remediation. Keep a running outage report and make sure the appropriate managers see it. Also clarify the support agreement and escalation path for the infrastructure team. If after-hours response is expected, there needs to be an explicit on-call arrangement or SLA; otherwise people may be blaming a team for work nobody is funded or scheduled to perform.
Reply to this answer
Before assuming every problem requires a major network upgrade, collect evidence from the logs. Check the wireless controller, RADIUS or authentication server, DHCP leases, client 802.1X logs, access-point placement, and endpoint power settings. Repeated re-authentication failures can come from client drivers, low-power settings, authentication timeouts, or a misconfigured NAC system. Multiple DHCP servers, MTU problems, or inconsistent configurations across sites are also worth investigating. The symptoms should be correlated with timestamps so the team can identify whether the wireless network, WAN, authentication service, or endpoints are actually failing.
Reply to this answer
Protect yourself by putting recommendations and rejected requests in writing, then stop taking personal responsibility for decisions you don’t control. Escalate professionally, retain the relevant records, and focus on the parts you can improve. If leadership repeatedly accepts the same outages and refuses reasonable fixes, you have three realistic choices: accept the environment, move to a company that invests in IT, or stay long enough to build experience without letting the situation consume you. Don’t care more about the infrastructure than the people who control its budget.
Reply to this answer
This is usually a management and prioritization problem rather than a technical mystery. The infrastructure team may have proposed fixes repeatedly, only to have them rejected because leadership decided the cost was higher than the expected downtime. IT’s job is often to document the risk, estimate the business impact, and price the solution; management then chooses whether to accept that risk. If the decision is never revisited, the business has effectively decided that the outages are acceptable.
QuietHarbor21: A company having plenty of revenue doesn’t necessarily mean it has approved budget for networking. The strongest case is a comparison of outage costs, affected staff, lost production, and recovery time against the cost of the proposed fix.
Reply to this answer
Add your answer
Answers are reviewed before publication.
BriskMeadow5: Extending the re-authentication interval might reduce the disruption temporarily, but it should be treated as a test or workaround, not a replacement for finding the root cause. Captures and client logs are much more useful than guessing.