The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Slow responses, refused connections, timeouts and error pages can all come from the same server, and they rarely say which layer is failing. The fastest route to a fix is to narrow the problem to one fault domain (host resources, storage, DNS or the network path, the application or service, or the operating system) and collect evidence before you change a configuration, restart a service or reboot the machine.
The procedures below follow Microsoft Learn guidance for Windows Server and AWS guidance for Linux instances on Amazon EC2. Treat the product-specific steps as examples for those platforms. They do not carry over unchanged to every Linux distribution, to other clouds or to physical servers, so confirm commands and menu paths against your own version before relying on them.
As an Amazon Associate I earn from qualifying purchases.
Start with the symptom, then work out the fault domain
The same complaint can come from several places. A slow application may be waiting on a saturated disk, a stalled DNS lookup or a service that has stopped answering. The table lists the fault domains each common symptom should prompt you to separate, and the first evidence worth collecting.
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Fix the driver behind crashes, sound loss and screen glitches3Repair Windows errors before they cause bigger problems| Reported symptom | Fault domains to separate | First evidence to collect | Platform guidance |
|---|---|---|---|
| Slow responses | Resource bottleneck across processor, memory, storage and network; application load; DNS delay | Counter or Linux measurements covering the same time window as the complaint, compared with a normal-period baseline | Windows Server: Performance Monitor. EC2 Linux: system log, iostat, iftop |
| Not responding or hung | Operating-system or kernel fault; memory exhaustion; storage or filesystem errors; a stopped service | Event log entries (Windows) or system log and status checks (EC2 Linux) | Windows Server: Event Viewer and Server Manager. EC2 Linux: instance status checks and system log |
| Cannot resolve the server name | DNS client configuration or connectivity; DNS server service, authoritative data, recursion or zone transfer | Client and server data captured at the same time while the failure is reproduced | Windows Server: Microsoft DNS troubleshooting guidance |
| Name resolves but connections fail | Network path or host reachability; application not listening on its port | Instance and system status checks; application status checks where configured | EC2 Linux: AWS status check documentation. Windows Server: not covered by the Microsoft guidance referenced here |
Scope the failure before changing anything
Record these five facts before you open a configuration file. They decide which evidence matters.
#1 Best Overall
- Dell PowerEdge R730xd 24B SFF 2U Server
- 2x Intel Xeon E5-2690 v4 2.6Ghz 14-Core (28-cores Total)
- 128GB DDR4 RAM – 4x 1.2TB 10K SAS 2.5” 12Gb/s
- Dell H730P mini 2GB 12Gb/s RAID
- 2x 750W PSU - 2x 10Gb SFP+ 2x 1Gb (RJ45) NIC
- What is broken. The whole host, one service, one application, DNS, or one client’s path to the server.
- Who is affected. All users, one site or subnet, or a single client. If only one client fails, its configuration is the first suspect.
- When it started, and what changed just before. Patches, configuration edits, deployments, DNS record changes, and resize or scaling events.
- Whether it is total or intermittent, and whether it is reproducible or load-dependent. A failure that appears only at peak load suggests a capacity limit; one that never recovers suggests a component that has stopped.
- Timestamps with their time zone. This lets event logs, performance counters and cloud metrics be lined up on one timeline.
Treat four questions as separate checks
Each of these needs its own test, because passing one does not clear the others:
- Is the application answering requests?
- Is the host reachable on the network?
- Does the server name resolve correctly from the affected client?
- Are processor, memory, storage and network resources within a normal range for the workload?
On EC2, AWS separates instance and system status checks, which describe the host, from application status checks, which can monitor network reachability and the availability of applications running on the instance.
Collect evidence before you change anything
The two platforms expose different evidence. Use the column that matches the server you are working on.
Rank #2
- Model: Dell OptiPlex 7050 Small Form Factor (SFF)
- Processor: Intel Core i7-7700 3.60 GHz
- Memory: 32GB DDR4 Ram
- Storage: 1TB Solid State Drive (SSD) Fast Boot + Storage
- Operating System: Windows 11 Pro (64-bit)
| Axis | Windows Server | EC2 Linux instance |
|---|---|---|
| Primary evidence | Event Viewer logs, service alerts in Server Manager, Performance Monitor counters, DNS audit and analytic logs, network traces | Instance and system status checks, system log, console output, CloudWatch metrics, Linux command-line tools |
| Fault domains the guidance covers | Resource bottlenecks, DNS resolution and server-side DNS configuration, service and event-log failures | Boot and reachability, memory, device, kernel and filesystem errors, operating-system configuration, network traffic and disk I/O |
| Operational risk of collecting evidence | Diagnostic logging and traces add load and use disk (see the DNS section) | Restarts and stops interrupt service |
| Access needed | Administrative access to the server, or remote access through Server Manager | Console access to instance status and logs, plus shell access if the instance still accepts logins |
Windows Server
Microsoft documents Server Manager’s event, performance and service monitoring for local and remote servers on Windows Server 2016, 2019, 2022 and 2025.
- Open Server Manager and select the affected server, local or remote.
- Go to Tools > Event Viewer, expand Windows Logs, and review System and Application for error and warning entries that begin close to the start time you recorded.
- Check the service alerts for any service that should be running, then confirm its state in the Services console (services.msc).
- Go to Tools > Performance Monitor (or run perfmon.msc) and add counters for the resources in question. Common choices are Processor > % Processor Time, Memory > Available MBytes, PhysicalDisk > Avg. Disk Queue Length, and Network Interface > Bytes Total/sec for each adapter. Record across a period that includes the incident and a normal period for comparison.
EC2 Linux instance
- In the EC2 console, open the instance and review its system status checks and instance status checks, plus any application status checks you have configured. Note which check fails and when.
- Open the system log with Actions > Monitor and troubleshoot > Get system log. Look for kernel, memory, device and filesystem messages around the failure time.
- If you can log in, read the kernel and system messages on the instance. On systemd-based distributions, journalctl -k shows kernel messages, and journalctl -b -1 shows the previous boot’s log after a restart.
- Review CloudWatch metrics for processor, network and disk activity across the incident window, and compare them with a normal period.
- For live traffic and disk activity, use iftop for network traffic and iostat for disk I/O. Neither may be installed by default; iostat is typically provided by the sysstat package.
Test the likely fault domain
Once you have scope and evidence, test the domain the evidence points to. Treat each result as a clue to confirm against the others.
Resource bottlenecks
Read processor, memory, storage and network measurements together, against a baseline from normal operation. A single high reading is a clue, not a diagnosis. High processor use can be a symptom of memory shortage, a storage queue or a saturated network link, so do not assume the processor is the cause.
Rank #3
- 2.80 GHz processor speed ensures efficient operation with consistent reliability
- Intel Xeon 2.80 GHz processor provides enterprise-grade performance with built-in security and remote management capabilities
- Quad-core (4 Core) processor core helps server process data quickly and reliably for maximum productivity
- 1 processors supported for faster processing and improved access to data, optimizing performance under heavy loads
- With 16 GB memory, you can multitask between applications seamlessly, keeping productivity high and response times quick
Microsoft’s Performance Monitor counter guidance, which carries a 2026 publication date, labels Network Interface Bytes Total/sec utilization below 50% as healthy, 50 to 80% as warning, and above 80% as critical. These bands belong to that one counter guide; they are not general server-health thresholds. The guide notes that interpretation depends on the adapter’s speed and the server’s role, so compare traffic with what that role normally sends and receives.
The same guide uses 8 bits = 1 byte when relating throughput units. Counters report bytes per second, while adapter speeds are quoted in bits per second, so convert before comparing. As an illustrative calculation, a 1 Gbps adapter has a line rate of 125,000,000 bytes per second, which puts the 50% and 80% points at 62,500,000 and 100,000,000 bytes per second. This is arithmetic, not a measured result from a server.
Storage and filesystem problems
A server can appear hung while its processor is idle when a volume is full, a disk returns I/O errors or a filesystem reports errors. Check the event logs (Windows) or the system log (EC2 Linux) for disk and filesystem errors, confirm free space on the affected volumes, and look for disk queue or wait figures that stay high: Avg. Disk Queue Length on Windows, and iostat output on Linux.
Rank #4
- MODEL P74439-005: Compact and affordable HPE ProLiant MicroServer Gen11 powered by Intel Pentium Gold G7400 3.7GHz processor, ideal for file sharing, NAS, and basic business workloads
- READY OUT OF THE BOX: Includes 16GB DDR5 UDIMM memory (expandable to 128GB), one 1TB SATA 6G Business Critical HDD, embedded Intel VROC SATA, dedicated iLO-M.2 port kit, 180w external power adapter and 1/1/1 warranty for dependable plug-and-play server operation
- WHISPER-QUIET & SPACE-SAVING: Ultra-compact mini tower design fits easily in small office spaces; supports wall, flat, or vertical placement for deployment flexibility
- INTEGRATED REMOTE MANAGEMENT: Comes with HPE iLO 6 and embedded TPM 2.0 for secure, license-free remote server administration through shared port access
- EXPANDABLE DESIGN: Two PCIe slots (including PCIe 5.0) and four LFF-NHP drive bays provide robust options for storage and component scalability. Features new MR408i-p controller support for enhanced storage performance
Application and service failures
When the host responds but the application does not, first confirm that the service is running. A stopped or crashed service usually leaves entries in the event log or system log at the time it stopped. Then test whether the service answers on its port from a second client, which separates a service fault from a client-side or path problem.
Boot, reachability and unresponsive instances
For boot failures or severe unresponsiveness on EC2 Linux, classify the failure before choosing a recovery action. AWS’s troubleshooting guidance groups example log problems into five categories:
Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstall- Memory: out-of-memory messages from the kernel.
- Device: block-device I/O errors.
- Kernel: kernel errors.
- Filesystem: filesystem errors.
- Operating-system configuration: problems in the system’s configuration.
These categories are examples from AWS documentation, not a complete taxonomy of server faults. If the log matches none of them, record the failure as unclassified and keep collecting evidence rather than forcing it into a category.
Best Value
- HP Z4 G4 Workstation Tower
- Intel Xeon W-2133 6-Core 3.6GHz (3.9GHz Turbo)
- 64GB DDR4 Memory - Nvidia Quadro P400 2GB
- 512GB NVMe M.2 SSD (boot) + 2TB HDD (storage)
- Windows 11 Pro 64-bit
Diagnose DNS problems
Microsoft’s DNS guidance for Windows Server separates client-side causes (IP configuration and connectivity on the client) from server-side causes (the server’s IP configuration, the DNS server service, authoritative data, recursion and zone transfer). It recommends starting on the client unless the scope already points to the server.
Start on the affected client
- Run ipconfig /all and confirm that the client’s DNS server address is the one you expect.
- Run nslookup for the server name and give the DNS server you expect as the second argument. This shows the answer that server returns, rather than whatever the client’s default resolver provides.
- Test basic connectivity to the server’s IP address with ping. If the address answers but the name does not resolve, DNS is the likelier fault. If neither answers, investigate connectivity first. A failed ping alone does not prove the host is down, because some networks filter ICMP.
Check the server side
| Check | Question it answers |
|---|---|
| IP configuration of the server | Does the DNS server have the correct address and network settings? |
| DNS server service | Is the service running, and does Event Viewer show errors from it at the time of the failure? |
| Authoritative data | Do the zone records for the name exist and contain the expected values? |
| Recursion | When a name lies outside the server’s own zones, does it resolve that name as designed? |
| Zone transfer | Are secondary servers receiving current zone data from the primary? |
Collect client and server data at the same time
Microsoft recommends collecting data on the client and the server simultaneously when feasible. Start the traces on both sides, reproduce the failure while they run, then stop and save the traces. A trace from one side alone cannot show whether a query reached the server or what the server answered.
Diagnostic DNS logging
Microsoft’s DNS logging guidance sets out the default states below and gives one performance example.
| DNS log | Default state (per Microsoft) | Cost and caution |
|---|---|---|
| Audit logs | Enabled by default | Not stated |
| Analytic logs | Not enabled by default | In Microsoft’s example on modern hardware, enabling analytic logging at 100,000 DNS queries per second caused about 5% performance degradation. The page reports no apparent impact at 50,000 queries per second and lower. These are examples from that page, not guarantees. |
| Debug logging | Not stated | Resource intensive and can consume disk space |
Disable analytic and debug logging once the trace is saved, and watch server performance while they are running.
Make a targeted change and verify it
No single remedy fixes every server problem, and Microsoft’s and AWS’s documentation do not describe one recovery sequence that applies across all of them. Once the failing component is identified, use the platform’s documented fix for that component, and wrap the change in the following control process.
- State the fault domain in one sentence and list the timestamped evidence that supports it.
- Confirm a current backup or rollback point exists for the component you are about to change. For production systems, follow your organisation’s change, backup and escalation procedures.
- Change one likely cause at a time where that is operationally possible. Record the old value, the new value and the time of the change.
- Reproduce the original failure and compare the same measurement over a comparable window. The fix is confirmed when the symptom and the measurement change together.
- If the symptom persists or the evidence points elsewhere, revert the change and return to the fault-domain step.
Avoid restarting, rebooting or replacing hardware as a first move. A restart ends active Performance Monitor sessions and open connections before you have captured them. Once the evidence is saved and the failure is classified, a restart becomes a reasonable next step.
Quick Recap
The Bottom Line
“”
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
