|
19 | 19 | \begin{frame}
|
20 | 20 | \frametitle{Devops, but Operations}
|
21 | 21 |
|
22 |
| -Services are doing The Thing -- but how do we know if there's a problem? |
| 22 | +Services are doing The Thing---but how do we know if there's a problem? |
23 | 23 |
|
24 | 24 | Monitoring and Alerting.
|
25 | 25 |
|
|
39 | 39 |
|
40 | 40 | Maybe automate testing the basic workflow.
|
41 | 41 |
|
42 |
| -These don't show performance problems. |
| 42 | +But, that doesn't show performance problems. |
43 | 43 |
|
44 | 44 | \end{frame}
|
45 | 45 |
|
|
108 | 108 | \includegraphics[width=0.8\textwidth]{images/more-sherlock.jpg}
|
109 | 109 | \end{center}
|
110 | 110 |
|
111 |
| -Remember also the lesson from Sherlock Holmes in ``The Adventure of Silver Blaze'' -- the dog that did \alert{not} bark was a clue as to who did the crime. |
| 111 | +Remember also the lesson from Sherlock Holmes in ``The Adventure of Silver Blaze''---the dog that did \alert{not} bark was a clue as to who did the crime. |
112 | 112 |
|
113 | 113 | \end{frame}
|
114 | 114 |
|
|
118 | 118 |
|
119 | 119 | The final option for detecting a problem is customer support.
|
120 | 120 |
|
121 |
| -Automated monitoring maybe can't find everything every time, but we shouldn't be relying on this as the primary mechanism. |
| 121 | +Automated monitoring maybe can't find everything every time, but we shouldn't be relying on reports from support as the primary mechanism. |
122 | 122 |
|
123 | 123 |
|
124 | 124 | \end{frame}
|
|
130 | 130 |
|
131 | 131 | \begin{itemize}
|
132 | 132 | \item {\bf Alerts}: a human must take action now;
|
133 |
| -\item {\bf Tickets}: a human must take action soon \\ \qquad (hours or days); |
| 133 | +\item {\bf Tickets}: a human must take action soon (hours or days); |
134 | 134 | \item {\bf Logging}: no need to look at this \\ \qquad except for forensic/diagnostic purposes.
|
135 | 135 | \end{itemize}
|
136 | 136 |
|
|
233 | 233 |
|
234 | 234 | Does not have to be exactly 5; use judgement.
|
235 | 235 |
|
236 |
| -Too few times $\rightarrow$ superficial answers. |
| 236 | +Too few times $\Rightarrow$ superficial answers. |
237 | 237 |
|
238 |
| -Too many times $\rightarrow$ ``computers were a mistake.'' |
| 238 | +Too many times $\Rightarrow$ ``computers were a mistake.'' |
239 | 239 |
|
240 | 240 | \end{frame}
|
241 | 241 |
|
|
275 | 275 | \item Irrelevant detail
|
276 | 276 | \item Speculation
|
277 | 277 | \item Blaming
|
278 |
| - \item Blaming |
| 278 | + \item Shaming |
279 | 279 | \end{itemize}
|
280 | 280 |
|
281 | 281 |
|
|
298 | 298 | \begin{frame}
|
299 | 299 | \frametitle{Risk Enormous}
|
300 | 300 |
|
301 |
| -You can run some program with tons of security vulnerabilities offline and feel that the security problems can be managed. |
| 301 | +You can run some program with tons of security vulnerabilities offline and feel that the security problems can be managed (but you might be wrong). |
302 | 302 |
|
303 |
| -When it's online the risk is enormous. |
| 303 | +When it's online the risk is definitely enormous. |
304 | 304 |
|
305 | 305 | All kinds of vulnerabilities are a problem, but I'll call out two of them:
|
306 | 306 |
|
|
327 | 327 |
|
328 | 328 | Bypass signup limitations, browser scripting.
|
329 | 329 |
|
330 |
| -Mining cryptocurrency -- inefficiently -- but works. |
| 330 | +Mining cryptocurrency---inefficiently---but works. |
331 | 331 |
|
332 | 332 |
|
333 | 333 | Using \$103~000 of resources produces one coin worth about \$137: 0.13\% return.\\
|
|
0 commit comments